Plant-Based Predictions: An Exploratory Predictive Analysis of Purchasing Behavior of Meat-Alternatives by U.S. Consumers (2020)

Despite the recent increase of plant-based diets, animal-based product consumption remains a major environmental concern. Present information science research is focused on the role of consumer perception in meat-alternative purchasing behavior and its marketing implications. This paper shifts from the profitability aspect of consumer behavior and seeks to understand how external variables such as urban residency, poverty, grocery store access, household income, food expenditure, and grocery costs relate to a county’s likelihood to purchase meat-alternatives. Using consumer survey responses from MRI Simmons and statistics from the U.S. government, we developed a logistic regression model, a support vector machine, and a generalized additive model to predict the likelihood of households in a county purchasing meat-alternatives. All features except for grocery access proved to be significant, positively correlated predictors. We conclude that features of physical and financial accessibility are useful in identifying, with roughly 68% accuracy, a U.S. county’s tendency for purchasing meat-alternatives. This identification might further sustainability initiatives and local efforts to incentivize environmentally conscious food decisions.