Measuring palatability as a linear combination of nutrient levels in food items

It well known that palatability and nutritional quality of foods and/or diets are viewed as being in tension with one another. While there exist multiple measures of healthiness, there are no such measures for tastiness. This gap limits the degree to which researchers can investigate this tension and its implications for dietary behavior and hence public health and nutrition policy. The scope of future work concerning the dietary behavior of Americans would expand greatly if researchers better understood consumers’ willingness to eat certain foods, which matters as much as recommending those foods for them to eat in the first place. Using stepwise selection algorithms, a nutrient profiling model is developed and the resulting estimates are used to compute a numerical measure for the relative palatability of food items. Foods found to be relatively tasty by the measure tend to be relatively lower in nutritional quality, as expected. This implies that policy aimed at altering consumption patterns should emphasize foods that score relatively high in both nutritional quality and tastiness. Tastier foods without the additional benefit of a healthier nutrient composition are candidates for consumption taxes to discourage excess consumption, while healthy foods low in palatability could be subsidized using this tax revenue.