Education was also found to significantly enhance label usage in a number of related
studies. Those with higher levels of education were found to be more likely to use
nutritional labeling information (Guthrie et al.; Bender and Derby; Feick, Herrmann, and
Warland; Klopp and MacDonald; Nayga). Although the variable for education was
estimated with a sign consistent to many previous studies, it was statistically
insignificant. Other insignificant variables included a dummy variable which denoted
the primary household shopper, income, and a variable which captured the effect of
households which made frequent usage of food advertisements and coupons for food
products.
The logit model chi-square statistic was significant at the 0.003 level clearly rejecting
the null hypothesis that the set of explanatory variables were together insignificant in
predicting variation in the dependant variable. The tabulation of prediction success is
shown in the classification table (Table 3). With a 50-50 classification scheme,
approximately 72 percent of the individuals in the sample were correctly classified as
those who place a high degree of importance on nutritional labeling when selecting
grocery products.
Conclusion
Qualitative choice models are ideal for analysis of many types of consumer behavior.
This study illustrates the potential of a logistic framework in decomposing the effects of
individual demographic characteristics in decision making. Other applications include
16