Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Technology-assisted Journal Writing for Improving Student Mental Wellbeing: Humanoid Robot vs. Voice Assistant

Batuhan Sayis, Hatice Gunes

TL;DR

The study investigates whether embodiment in agent-assisted journaling (humanoid robot Pepper vs. voice assistant Google Nest Mini) affects student mental wellbeing. Using a 2x2x2 mixed factorial design across four weekly sessions (Q vs. Q+J; SR vs. VA; week1 vs. week2), it measures mood (BMIS), self-disclosure (SSDQ), and agent perception (RoSAS). Results show robot-assisted journaling yields mood improvements, higher disclosure in the form of awareness of thoughts/feelings, and reduced discomfort over time, whereas the voice assistant shows no significant effects. These findings suggest robot embodiment can enhance the therapeutic potential of technology-assisted journaling for student wellbeing, with implications for the design of mental health support tools in educational settings.

Abstract

Conversational agents have a potential in improving student mental wellbeing while assisting them in self-disclosure activities such as journalling. Their embodiment might have an effect on what students disclose, and how they disclose this, and students overall adherence to the disclosure activity. However, the effect of embodiment in the context of agent assisted journal writing has not been studied. Therefore, this study aims to investigate the viability of using social robots (SR) and voice assistants (VA) for eliciting rich disclosures in journal writing that contributes to mental health status improvement in students over time. Forty two undergraduate and graduate students participated in the study that assessed the mood changes (via Brief Mood Introspection Scale, BMIS), level of subjective self-disclosure (via Subjective Self-Disclosure Questionnaire, SSDQ), and perceptions toward the agents (via Robot Social Attributes Scale, RoSAS) with and without agent (SR or VA) assisted journal writing. Results suggest that only in robot condition there are mood improvements, higher levels of disclosure, and positive perceptions over time in technology-assisted journal writing. Our results suggest that robot assisted journal writing has some advantages over voice assistant one for eliciting rich disclosures that contributes to mental health status improvement in students over time.

Technology-assisted Journal Writing for Improving Student Mental Wellbeing: Humanoid Robot vs. Voice Assistant

TL;DR

The study investigates whether embodiment in agent-assisted journaling (humanoid robot Pepper vs. voice assistant Google Nest Mini) affects student mental wellbeing. Using a 2x2x2 mixed factorial design across four weekly sessions (Q vs. Q+J; SR vs. VA; week1 vs. week2), it measures mood (BMIS), self-disclosure (SSDQ), and agent perception (RoSAS). Results show robot-assisted journaling yields mood improvements, higher disclosure in the form of awareness of thoughts/feelings, and reduced discomfort over time, whereas the voice assistant shows no significant effects. These findings suggest robot embodiment can enhance the therapeutic potential of technology-assisted journaling for student wellbeing, with implications for the design of mental health support tools in educational settings.

Abstract

Conversational agents have a potential in improving student mental wellbeing while assisting them in self-disclosure activities such as journalling. Their embodiment might have an effect on what students disclose, and how they disclose this, and students overall adherence to the disclosure activity. However, the effect of embodiment in the context of agent assisted journal writing has not been studied. Therefore, this study aims to investigate the viability of using social robots (SR) and voice assistants (VA) for eliciting rich disclosures in journal writing that contributes to mental health status improvement in students over time. Forty two undergraduate and graduate students participated in the study that assessed the mood changes (via Brief Mood Introspection Scale, BMIS), level of subjective self-disclosure (via Subjective Self-Disclosure Questionnaire, SSDQ), and perceptions toward the agents (via Robot Social Attributes Scale, RoSAS) with and without agent (SR or VA) assisted journal writing. Results suggest that only in robot condition there are mood improvements, higher levels of disclosure, and positive perceptions over time in technology-assisted journal writing. Our results suggest that robot assisted journal writing has some advantages over voice assistant one for eliciting rich disclosures that contributes to mental health status improvement in students over time.
Paper Structure (11 sections, 4 figures)

This paper contains 11 sections, 4 figures.

Table of Contents

  1. Introduction
  2. METHODS
  3. Participants & Study Design
  4. Agents Used
  5. Agent assisted journal writing (J)
  6. Agent assisted questionnaire responding (Q)
  7. Experimental Protocol
  8. Data Analysis
  9. RESULTS
  10. DISCUSSION
  11. LIMITATIONS & FUTURE WORK

Figures (4)

  • Figure 1: Diagram of the experimental design. Abbreviations: GROUP1 (SR), Robot condition; GROUP2 (VA), Voice Assistant condition; BMIS, Brief Mood Introspection Scale; PANAS, Positive and Negative Affect Schedule; SDHS, The Short Depression-Happiness Scale; RoSAS, Robot Social Attributes Scale; SSDQ, Subjective Self-Disclosure Questionnaire
  • Figure 2: Mean BMIS Negative-Calm sub-scale levels pre & post, in Q & Q+J, in week1 (w1) & week2 (w2). Lower values indicate less negative mood.
  • Figure 3: Mean SSDQ sub-scale levels, after Q+J, in week1 (w1) & week (w2). Only in SR, Q+J shows significant increase from w1 to w2 in the awareness of thoughts and/or feelings.
  • Figure 4: Mean RoSAS sub-scale levels, after Q & Q+J, in week (w1) & week (w2). Only in SR, Q+J shows significant decrease from w1 to w2 in the sub-scale of discomfort.