| dc.description.abstract | Eating healthy is crucial to maintaining good health and avoiding diseases. Within the technology field, many nutrition applications and chatbots have been developed to promote healthy eating. However, the latter still faces a few challenges regarding user engagement. Users of low Health Literacy tend to be less engaged in using health chatbots as they do not understand medical terms that well. Moreover, there is little research about the Nutrition and Food Literacy (N&FL) of the user. In our study, we assessed the N&FL of the user with the Nutrition Literacy Scale (NLS) and developed a nutrition chatbot. The chatbot adapted its text on numeracy, vagueness and wording and provided nutrition re- ports. We conducted a within-subject experiment with 22 participants divided into two groups, adaptive and non-adaptive, that interacted with the equivalent chatbot.
We measured user engagement using quantitative methods. No significant differences were found between the groups as all participants followed similar engagement behaviour. However, this increased when push notifications were sent as reminders. In addition, we conducted seven interviews to measure user satisfaction and if they gained any new knowledge. The adaptive group was identified with a higher interest in nutrition than the non-adaptive, leading to lower satisfaction because of higher expectations from the chatbot. No one expressed any knowledge gain, but rather they gathered more insights into their daily diet. Finally, the textual content analysis showed that each group had different interests in their questions. Although our experiment did not have significant results, we believe this study gives new insights into a topic that has not been studied as much, namely adapting a nutrition chatbot to the user’s interests. We suggest that future work focuses on this direction. Due to the limitations of our study, we also propose repetition of the experiment with a more diverse sample in N&FL and a different N&FL assessment tool. | |
