Answers are grounded in this paper and cite their sources.
Introduction Pregnancy is a health-constrained phase of life in which food-related decisions are shaped not only by individual preferences, but also by medical advice, family support, perceived safety, and digital information environments. In this context, green food is often discussed not merely as an environmentally preferable option, but as a potentially safer and higher-quality choice during pregnancy. Aims This study compares China and Thailand to examine how different forms of digitally expressed knowledge are configured into pathways associated with strong green food purchase orientations in pregnancy-related social media discourse. Methods Based on social media platform data, this study first used Python to collect 63,120 Chinese and 4,805 Thai public comments. Latent Dirichlet Allocation (LDA) was then applied for thematic analysis, followed by fuzzy-set Qualitative Comparative Analysis (fsQCA) to identify knowledge configurations associated with high green food purchase orientation in both Chinese and Thai samples, while also validating reverse pathways and causal asymmetries. Results We extract six knowledge dimensions through natural language processing (NLP) and latent dirichlet allocation topic modeling (1) Nutritional and risk knowledge, (2) Eco-label and certification recognition, (3) Food safety and institutional knowledge, (4) Digital health literacy and platform literacy, (5) Price-value conversion knowledge, (6) Social support knowledge. Results reveal that knowledge influences behavior through “configurational effects” rather than linear main effects. In Chinese, the pathway centered on institutional trust and label recognition was most significant, whereas in Thai samples, a platform-oriented, price-sensitive pathway structure emerged. Both country-specific pathways demonstrated causal asymmetry, with distinct positive and negative pathways. Implications The findings suggest that knowledge does not operate as an isolated linear driver; rather, digitally expressed knowledge becomes influential when combined with trust, affordability, and social support in specific configurations.
No takes yet. Share an insight, caveat, or question.
Huang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: