Title: Using explainable food swaps to nudge users towards more sustainable products in grocery websites
University: University of Bergen (UiB), Department of Information Science and Media Studies
Supervisor: Prof. Df. Christoph Trattner Co-supervisor: Dr. Alain D. Starke
The growing concern for climate issues har prompted both consumers and the grocery retail industry to prioritize environmental sustainability. This thesis aims to examine the effectiveness of nudging usrs towards more sustainable food options in an online grocery store using food swaps. Further, the study utilizes different motivational explanations accompanying the swaps, to investigate their impact on swap acceptance and perceived understanding. A mockup supermakret interface was created and screenshots uploaded to an online survey tool, where participants (N=202) were assigned to one of four conditions (baseline, health, sustainability or money). Results indicate that motivational framing did not significantly influence swap acceptance. However, perceived understandability was significant in affecting swap acceptance, with the sustainability framing being better understood. Participants were more likely to swap when the cost of the alternative product increased, suggesting other factors influenced consumer behavior. Finally, perceived similarity between the original and alternative product significantly affected the swap acceptance and perceived similarity, where meat swaps showed a strong, positive statistical significance. This thesis provided novel work within the field of encouraging more sustainable products in online grocery shopping services, which can be further expanded by implementing sustainable food swaps in a recommender system.
The repository includes the prototype, with the HTML, JavaScript and CSS code that make up the mockup supermarket interface. Further, it includes the dataset, which is utilized in the prototype. In addition, the folder "Web crawler" obtain the Python code used to collect the dataset. Lastly, the folder "Data analysis" contain the user data from the study, both cleaned and unprocessed, an excel sheet with some of the descriptive statistics and the R script used to analyze the data.