Keywords
chatbot, psychobio, usability, HCI, consent

Technologies
React, Node.js, SQLite, Hugging Face Transformers

Psychobiological research often requires participants to collect biological samples at home, without a researcher present. This can be uncomfortable, both practically and emotionally, and that discomfort is a real barrier to participation. The broader project aimed to design and develop a chatbot onboarding system and compare its effectiveness to human researcher onboarding. This dissertation focused specifically on the usability and user experience of that system. Usability was tested with 30 participants using the Chatbot Usability Questionnaire, interaction logs, and open-ended feedback. The system scored 65.1 out of 100, just under the average benchmark of 68. Participants liked the step-by-step layout, progress bar, and scroll-back functionality. Many also noted that talking to a chatbot felt less embarrassing than going through the same process with a researcher, which matters for a sensitive procedure like this. The biggest problem was response speed, averaging 37.4 seconds per reply, which broke the flow and in one case clashed with a time-sensitive instruction. Overall the chatbot showed real promise, but better response quality and faster load times are needed.