In online questionnaires, free-text inputs and open-ended questions often receive sparse or unrelated responses, even when participants are compensated for their participation. This thesis aims to explore how participants can be nudged into providing more meaningful answers. Approaches could include simplistic implementations such as icon feedback based on text length or more advanced feedback techniques backed by a Large Language Model (LLM) to offer real-time feedback on the quality of their responses. The thesis will involve the development of "nudging-input-fields" that integrate LLM feedback and a study to compare their effectiveness against traditional input methods in both paid and unpaid participation scenarios.

Focus

• Develop a questionnaire prototype using web technologies that integrates different real-time feedback methods on input fields.
• Conduct a study to evaluate the effectiveness of these nudging-input-fields in improving the quality of participant responses.
• Analyze the results by comparing the nudging-input-fields to classical input methods across different compensation models.

Prerequisites

• Experience as a user with Large Language Models (LLMs) such as ChatGPT or Gemini.
• Strong web development skills, including both frontend and backend.
• Familiarity with HCI principles and study design is a plus.

How to Apply

Please send an email with the following information to nina.knieriemen@dfki.de (incomplete applications will not be considered):

1. Planned start date for the thesis
2. Planned completion date for the thesis
3. A short motivational statement explaining why this topic interests you
4. A summary of why you are a good fit for this topic, especially outlining your prior experiences that will help you in successfully working on this topic.
5. Your transcript of records and CV
6. A one-page document describing which types of nudging-input-fields you think are effective and an explanation why