Data from a national survey of food shoppers are analyzed by probit and ordered probit models that incorporate elements of Lancaster's product attribute model and Weinstein's precaution adoption process. The models are used to investigate the characteristics of organic and non-organic food shoppers. Where one shops, food beliefs and food knowledge have the largest significant impact on the probability that shoppers buy organic food. Among the demographic characteristics, only the lack of religious affiliation, higher education, and youth are significant explanatory variables.