Reflection can be defined as ”mechanism to learn through experience“ [1]. Wolfbauer et al. have shown in [2] that reflection is possible in trainees with the help of a chatbot and a reflection script. This master thesis deals with the question whether a successful application of the reflection script with a chatbot is transferable to other domains. This is tested in a use case in a medical context with residents of the CharitÅLe Berlin, where the reflection script from [2] is additionally contrasted with a control group with a general reflection prompt [3][4] and the number of reflective utterances is compared. In addition, a second use case tests the extent to which reflection is possible independent of domain knowledge. Here, the control group also receives a general reflection prompt.

The result of the two studies is that a group that reflects using the reflection script fromWolfbauer [2] makes an average of 4.27 reflective utterances – coded using reflection codes according to [5] – makes, while a comparison group with a general reflection prompt achieves 2.84 reflective utterances. No correlation with the number of reflective utterances was found with prior knowledge of the object of reflection – as measured by the Security Behavior Intention Scale [6] – nor with educational attainment. Interaction with the chatbot was perceived as positive. Further development of the reflection script with the support of a chatbot, for example adaptively according to the user’s needs and expressions, may therefore be promising.

References

[1] Birgit R. Krogstie, Michael Prilla, and Viktoria Pammer. Understandingand Supporting Reflective Learning Processes in the Workplace: The CS-RL Model. In Davinia Hern ́andez-Leo, Tobias Ley, Ralf Klamma, andAndreas Harrer, editors,Scaling up Learning for Sustained Impact, volu-me 8095 ofLecture Notes in Computer Science, pages 151–164. SpringerBerlin Heidelberg, 2013.

[2] Irmtraud Wolfbauer, Viktoria Pammer-Schindler, and Carolyn P Rose.Rebo Junior: Analysis of Dialogue Structure Quality for a Reflection Gui-dance Chatbot. page 14, 2020.

[3] Elizabeth A. Davis. Prompting Middle School Science Students for Pro-ductive Reflection: Generic and Directed Prompts.Journal of the LearningSciences, 12(1):91–142, 2003.

[4] Dirk Ifenthaler. Determining the effectiveness of prompts for self-regulatedlearning in problem-solving scenarios.Journal of Educational Technology& Society, 15(1):38–52, 2012.

[5] Serge Egelman and Eyal Peer. Scaling the Security Wall: Developinga Security Behavior Intentions Scale (SeBIS). InProceedings of the 33rdAnnual ACM Conference on Human Factors in Computing Systems, pages2873–2882. ACM, 2015.