Caloric soft drinks are the number one source of added sugars in U.S. diets, and are associated with many health problems. Three recent years of household purchase, household demographic, and industry advertising data allow Heckit estimation to identify how specific demographic groups vary in their purchase response to marketing of sweetened carbonated soft drinks (sCSDs) at the product category level. Empirical results reveal unique non-linear patterns of household purchase response to sCSD-industry price, sale, and advertising signals that vary significantly by specific demographic characteristics. Isolating the effects of either price, sale, or advertising on household purchase, highest education level of high school or less for the household head tends to be the most robust predictor of higher sCSD purchase, followed by household income at or below the poverty level for a family of four. The novel approach and results here contribute to the literature by estimating how rising education level for a fixed level of household income will variously affect sCSD purchase quantity depending on the ethnicity of the household, and does the same fixing education level across rising income level. Econometric controls are used to avoid estimation and inference errors the literature warns commonly accompany the Heckman specification.