To analyze fresh vegetable consumption using household survey data, the tobit model and a more flexible parameterization to the tobit model—the “double hurdle” model—were considered. Based on the likelihood ratio test, the tobit model was rejected against the “double hurdle” specification. Moreover, the results suggest that the tobit model underestimated the impact of the explanatory variables on fresh vegetable expenditures. Other results indicate that total food expenditures (proxy for income), age, household composition, sex, race, marital status, urbanization, region, and seasonality are all important determinants of fresh vegetable expenditures.
