Chatbot-Driven Surveys in Longitudinal Health-Related Studies

Pawel Chwedoruk & Alexis Joseph Tschopp

Student thesis: Master thesis


Nowadays, the digitalisation of healthcare industry leads pharmaceutical companies, clinics and other health-related entities to administer multiple web-based surveys to patients. With the advent of a new mode of survey, chatbot-driven, and prior research on the negative relationship between multiple web-based survey requests and response rates, the following question arises: How do chatbot-driven surveys perform in longitudinal health-related studies in regard to response rates compared to webbased surveys? An exploratory experiment was designed based on a deductive approach: A longitudinal study on alternative treatments for autoimmune diseases was administered to 808 participants, equally split into four groups, each of which received four surveys in a two-week timespan. The mode of survey was either chatbot-driven or web-based, and the order of surveys was either 1-2-3-4 or 4-2-3-1; Surveys number 4 being sent at the same date and time; Facebook Messenger and Google Form being used for chatbot-driven and web-based surveys respectively. A quantitative data analysis demonstrated that chatbot-driven surveys did not only yield lower response rates than web-based surveys but were also more vulnerable to the impact of multiple survey requests on response rates. Based on the specific experiment setting, the findings suggest to reconsider the adoption of chatbot-driven surveys in health-related longitudinal studies. Further research on the generalisability of these results and the nature of the aforementioned relationships is highly recommended

EducationsMSc in Business Administration and Information Systems, (Graduate Programme) Final Thesis
Publication date2018
Number of pages178
SupervisorsAbayomi Baiyere