Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DiaryHelper: Exploring the Use of an Automatic Contextual Information Recording Agent for Elicitation Diary Study

Junze Li, Changyang He, Jiaxiong Hu, Boyang Jia, Alon Halevy, Xiaojuan Ma

TL;DR

This paper addresses recall bias and diary burden in elicitation diary studies by introducing DiaryHelper, an AI agent that predicts five episodic-context cues (time, location, emotion, people, and activity) for each entry, leveraging LLMs and multimodal cloud APIs. In a within-subject two-week study with twelve participants conducted via Slack across four modalities, DiaryHelper reduced recording burden and increased the richness of recall and discussion in elicitation interviews. The results show high accuracy for emotion, people, and location predictions, with significant gains in recall for location and people, and overall improved elicitation quality. The work offers design guidance for topic-specific customization and demonstrates the potential of AI-generated contextual cues to enhance qualitative data collection in diary-based research.

Abstract

Elicitation diary studies, a type of qualitative, longitudinal research method, involve participants to self-report aspects of events of interest at their occurrences as memory cues for providing details and insights during post-study interviews. However, due to time constraints and lack of motivation, participants' diary entries may be vague or incomplete, impairing their later recall. To address this challenge, we designed an automatic contextual information recording agent, DiaryHelper, based on the theory of episodic memory. DiaryHelper can predict five dimensions of contextual information and confirm with participants. We evaluated the use of DiaryHelper in both the recording period and the elicitation interview through a within-subject study (N=12) over a period of two weeks. Our results demonstrated that DiaryHelper can assist participants in capturing abundant and accurate contextual information without significant burden, leading to a more detailed recall of recorded events and providing greater insights.

DiaryHelper: Exploring the Use of an Automatic Contextual Information Recording Agent for Elicitation Diary Study

TL;DR

This paper addresses recall bias and diary burden in elicitation diary studies by introducing DiaryHelper, an AI agent that predicts five episodic-context cues (time, location, emotion, people, and activity) for each entry, leveraging LLMs and multimodal cloud APIs. In a within-subject two-week study with twelve participants conducted via Slack across four modalities, DiaryHelper reduced recording burden and increased the richness of recall and discussion in elicitation interviews. The results show high accuracy for emotion, people, and location predictions, with significant gains in recall for location and people, and overall improved elicitation quality. The work offers design guidance for topic-specific customization and demonstrates the potential of AI-generated contextual cues to enhance qualitative data collection in diary-based research.

Abstract

Elicitation diary studies, a type of qualitative, longitudinal research method, involve participants to self-report aspects of events of interest at their occurrences as memory cues for providing details and insights during post-study interviews. However, due to time constraints and lack of motivation, participants' diary entries may be vague or incomplete, impairing their later recall. To address this challenge, we designed an automatic contextual information recording agent, DiaryHelper, based on the theory of episodic memory. DiaryHelper can predict five dimensions of contextual information and confirm with participants. We evaluated the use of DiaryHelper in both the recording period and the elicitation interview through a within-subject study (N=12) over a period of two weeks. Our results demonstrated that DiaryHelper can assist participants in capturing abundant and accurate contextual information without significant burden, leading to a more detailed recall of recorded events and providing greater insights.
Paper Structure (38 sections, 6 figures, 11 tables)

This paper contains 38 sections, 6 figures, 11 tables.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Diary Study Implementations
  4. Challenges and Difficulties in Diary Study
  5. Improved Diary Study Methods
  6. Techniques for Facilitating Recall
  7. System Design
  8. Selection of Contextual Information
  9. Implementation
  10. Front-End Integration with Slack
  11. Multimodal Data Processing
  12. Contextual Information Prediction by LLM
  13. User Interface
  14. Evaluation
  15. Study Procedures
  16. ...and 23 more sections

Figures (6)

  • Figure 1: The interface of DiaryHelper .
  • Figure 2: The general pipeline of our study (participant group 1/2 are denoted as G1/2).
  • Figure 3: Samples of recorded diaries in different modalities.
  • Figure 4: The number of participants' recorded diary entries per hour during a day.
  • Figure 5: Participants' ratings of DiaryHelper regarding satisfaction, carefulness and privacy concern.
  • ...and 1 more figures