Research on College Students’ Willingness to Pay a Premium for Degradable Takeout Packaging Based on Random Forest-SHAP Interpretable Model

Abstract

Takeout packaging pollution on university campuses remains a severe issue, hindering the popularization of degradable packaging. Against this backdrop, this study takes college students as the research object and collects 600 valid samples via stratified random sampling. Integrating the Theory of Planned Behavior and the Norm Activation Theory, this paper defines environmental protection behavioral belief and purchase intention as chained mediating variables. Five machine learning models including linear regression are constructed and compared; the random forest model with optimal fitting performance is selected, and SHAP interpretable analysis is adopted to rank feature importance and screen independent variables. Ultimately, a research model is established, which takes green packaging knowledge, personal norms and degradable packaging quality as independent variables, environmental protection behavioral belief and purchase intention as chained mediators, and acceptance of degradable packaging and premium payment behavior as dependent variables. Regression analysis and chained mediation tests are conducted. The results show that green packaging knowledge, personal norms and degradable packaging quality can directly promote college students’ premium payment behavior, and also exert indirect positive impacts on premium payment behavior through the chained mediating path of “environmental protection behavioral belief → purchase intention”.