While causations are mostly discovered by quantitative analysis, the causality in qualitative data for social scientists is also crucial for understanding local contextualized events and processes, yet it requires going beyond simple self-reports and necessitates the careful analysis of both explicit self-reported reasoning and tacit causation patterns that emerge from a chronological examination of events and behaviors.
To enhance this analytical process, we present an interactive visualization system that illuminates causal relationships within qualitative data. Through network analysis and interactive visualization, our system aims to transform qualitative data into interpretable visual patterns that reveal underlying causal mechanisms. This website hosts the “live” version of our system, beginning with Version 1.0, and will be continuously updated to incorporate new features and improvements. Our inaugural release features a case study on mental-illness stigma. In this version, we aim to help social scientists understand the causal mechanisms underlying stigma formation and attributional reasoning.
Welcome to our project page! Check our system here: Interactive Visual Analytics System
Our qualitative data is derived from human-chatbot interactions based on the attribution model. We mainly analyze seven attributions: responsibility, anger, pity, fear, coercive segregation, social distance, and helping. We've designed seven corresponding question-answer pairs to address each attribution. For more information about our data acquisition, please refer to our preprint(s).
Small Talk
Chatbot: Hi, I’m Nova, nice to meet you! Could you please tell me your nickname? (Please just input your nickname)
User: Cynthia.
Chatbot: Hi Cynthia, nice to meet you! Welcome to our conversation.
Chatbot: How was your mood today?
User: Pretty good.
Chatbot: I’m so glad to know! / I’m sorry to hear that, hang in there!
Chatbot: What is your favorite thing to do after get off work or school?
User: Go out and exercise.
Chatbot: Exercise is a great way to unwind and stay healthy.
Chatbot: Do you watch movie recently? How about recommend one to me?
User: No actually. But I like The First Slam Dunk very much, and I’d like to recommend this to you.
Chatbot: Thanks for the recommendation! I'll check out "The First Slam Dunk”.
Chatbot: Watching movie can be really relaxing, have you ever attended a friend party? Do you like it?
User: yes, i like it
Chatbot: That's great to hear! Friend parties can be a lot of fun. Thank you for sharing!
Chatbot: Friend party might be enjoyable, hmm, but what do you think about a friend's party with many of your friend's friends, people you're not familiar with? Do you feel anxious or happy?
User: i think it’s ok for me, but i dont think i will talk to them haha
Chatbot: That's totally fine. Enjoying the atmosphere without engaging in conversation can be a great way to socialize!
Chatbot: As for me, it depends on the stranger’s personality. I can maybe learn something new from them, but also I may feel stressful if the person is not easy-going.
Vignette Delivery
Chatbot: Okay! I've got a story about Avery's adventure at a friend's party. Let's dive right in!
User: ok
Chatbot: Okay, I'll provide the vignette now.
Chatbot: Most of Avery's friends think that Avery is an easygoing and happy person, but this view has changed recently.
Chatbot: When Avery is at a company gathering, they start to isolate themselves and become irritable.
Chatbot: Avery also feels this change: they feel angry about their surroundings, and they get frustrated about where the fury comes from.
Chatbot: When they are asked by one of their colleagues to share their thoughts on the news, Avery loses their temper and jumps out of their seat.
Chatbot: They scream, “I hate this conversation!” as they run out of the place.
User: Oh..
Chatbot: Avery cannot explain why they feel the strong urge to do that.
Chatbot: They unexpectedly lose their temper in these instances, causing a sense of not fitting in with their friends.
Chatbot: Cynthia, have you ever had a similar feeling to Avery - feeling angry at a social event?
User: No
Chatbot: Understood. Thanks for sharing your experience.
Chatbot: Let me tell you more about Avery.
Chatbot: Avery finds that when they are depressed, they can barely concentrate on their work.
Chatbot: When Avery is working on their company projects, their work performance has declined, but they won't ask for help and feel like a burden.
Chatbot: They are considering quitting, feeling like a failure, and Avery's manager asks them to improve.
Chatbot: They have noticed that the online courses they initially enrolled in out of interest no longer excite them, so they gave them up.
User: Oh okay..
Chatbot: At home, Avery lacks motivation and neglects household chores, causing their family members frustration.
Chatbot: Avery feels guilty and empty. They feel that their family can't comprehend their suffering.
Chatbot: When interacting with strangers, Avery believes that others judge them harshly and they struggle to engage comfortably in conversations.
Chatbot: Every time they check Instagram and Facebook, Avery sees their friends are having fun, and Avery feels alone because they do not understand why they cannot be like that.
User: I see..
Chatbot: Hmmmm, the story is over. After reading it, what suggestions would you give to Avery?
User: I think he should go to hospital and have a check.
Chatbot: That's a reasonable suggestion, seeking professional help can provide clarity and support for Avery.
Key Questions Delivery - Q1
Chatbot: Cool! Let's talk more about the story. Just like what is mentioned in the story, Avery is currently facing difficulties in both their relationships with colleagues and their work performance. Do you believe Avery’s current situation is primarily a result of their actions?
User: no
Chatbot: I understand your perspective. Avery's situation may have other factors at play. Thank you for sharing your view.
Chatbot: Could you provide some reasons?
User: He is just ill, and it will be better when she recovered.
Chatbot: I understand your perspective. You believe Avery's situation is primarily due to his illness. Can you share more about why you think it's not his actions but his illness that's causing these issues?
User: His action is due to his illness. He has no fault, he even tried to become better.
Chatbot: I agree, his actions are driven by his illness, and he's actively seeking improvement.
Chatbot: For me, I think Avery’s current situation is a complex interplay of their actions and their mental health challenges. Their experience with mental illness does have an impact on their behavior to a certain extent, but their behavior has certainly contributed to strained relationships and reduced work performance.
Key Questions Delivery - Q2
Chatbot: What's more, may I kindly ask, if you were in the process of selecting a tenant for your home, would you feel comfortable entrusting it to someone like Avery?
User: I probably not
Chatbot: I get it, you're not keen on having Avery as a roommate. Thanks for sharing your preference.
Chatbot: May I ask why was it?
User: Because I’m afraid I will harm him, like conflict usually occur between roommates.
Chatbot: You mentioned concerns about potential conflicts with Avery as a roommate. Could you elaborate on those concerns and how they relate to his condition?
User: because everyone’s time is not the same. I get up very late, so if he get up early and make sounds, i will probably cannot control my mood. this will hurt him.
Chatbot: I understand your concern about potential conflicts due to differing schedules and the impact on Avery's well-being.
Chatbot: I might be open to the idea of renting the room to Avery, but I do hesitate because I'm worried about how their mental condition might affect aspects like cleanliness or the overall rental situation.
Key Questions Delivery - Q3
Chatbot: Thank you for sharing! Let's imagine you were one of Avery’s friends at the party or other social occasions. Would you feel anger toward them due to their actions, for example, losing their temper and yelling at someone?
User: maybe I will not
Chatbot: It seems you wouldn't feel angry with Avery in that situation. Your perspective is understood, thank you for sharing.
Chatbot: May I ask you why?
User: Because I think we can still have fun without Avery.
Chatbot: You mentioned that you might not feel angry because you think the party can still be enjoyable without Avery. Could you share more about why you believe the party can go on smoothly even without his presence?
User: We still have some friends on the party, so his left is not a big thing to us maybe.
Chatbot: I understand, having other friends at the party can make it enjoyable even if she leaves.
Chatbot: If I were one of Avery’s friends at the party or other social occasions, I might feel frustrated or annoyed because their actions disrupt social events. However, I would also understand that their behavior is likely due to their mental health challenges, which would temper my anger.
Key Questions Delivery - Q4
Chatbot: Thanks for the discussion with me!! I wonder if you were one of Avery’s colleagues or friends, then would you be inclined to help them with tasks, such as work projects?
User: no..
Chatbot: You wouldn't want to team up with him for work projects. Thanks for sharing your view.
Chatbot: How about your reason? Why?
User: Because it’s important to be responsible for myself’s work in my workplace. I want to show my ability, so I will choose a collaborator who is very strong. But no discrimination here..
Chatbot: I understand your perspective on preferring a strong collaborator for work projects. It's important to focus on productivity and teamwork.
Chatbot: For me, to be honest, I would consider giving them an opportunity to collaborate. I might also hesitate to cooperate with Avery because their condition could potentially impact their performance or reliability in the workplace. Hmmm, hard to choose.
Mid-break
Chatbot: Okay, let's take a brief break from the questions! Do you have any hobbies or interests that help you relax and unwind after work discussion?
User: Sleeping
Chatbot: Sounds refreshing! Getting a good night's sleep is essential for unwinding. 😴
Chatbot: Great! Thank you so much, Cynthia! I hope you do not feel tired now, let’s go back to our interesting discussion! Let me quickly recap of Avery's story:
Chatbot: Avery, once perceived as easygoing, is undergoing a transformation. At company gatherings, they isolate themselves, succumb to irritability, and unexpectedly explode with anger, leaving them feeling disconnected from colleagues. Their depression manifests in declining work performance, a reluctance to seek help and contemplation of quitting. Their once-enjoyed pursuits lose their appeal, contributing to a pervasive sense of guilt and emptiness. Interactions become strained, as Avery perceives harsh judgment from others. The contrast of joyful social media posts intensifies their feelings of loneliness and incomprehension.
Key Questions Delivery - Q5
Chatbot: I’m really happy you are with me now. Let’s continue our talk. I’d like to know if you would feel concern and sympathy for Avery’s mental state?
User: Yes sure..
Chatbot: Certainly, you would feel concern and pity for Avery. Your empathy is appreciated.
Chatbot: Let's go further. Is there any reason why you think this way?
User: ok. I will feel concern about Avery, because he is experiencing illness.
Chatbot: Thank you for sharing that. Can you elaborate on why you feel concerned when someone is experiencing illness?
User: because I feel he needs help, and his life is affected. It’s natural for us to feel synpathy on those people who are suffered.
Chatbot: I completely understand, and it's quite compassionate to want to help someone in need.
Chatbot: I’m more than glad to share my thoughts! In this scenario, I could have conflicting emotions regarding Avery’s situation. On one hand, I could feel concerned for them because their condition appears to be causing significant distress and affecting their daily life. On the other hand, I might also consider their condition as vulnerabilities that they should overcome, which could lead to a lack of sympathy.
Key Questions Delivery - Q6
Chatbot: Apart from being a colleague or friend, I’m also curious if you are Avery’s neighbor, then would you consider it better to separate them from the neighborhood and have them undergo hospitalization for treatment?
User: somehow yes
Chatbot: I appreciate your honesty. You see the condition as possibly indicating weakness. Thank you for sharing your perspective.
Chatbot: What’s your reason?
User: I consider normal people can do better in social event. But still, I think it’s because his illness, his illness lead to his weakness. it can be understand.
Chatbot: I understand your point. You believe that his illness contributes to his perceived weakness in social situations, which is understandable.
Chatbot: Well, I would probably think that placing Avery in a hospital can be an option. This could somewhat reduce the potential risk of having problems with the neighbors, and perhaps, without Avery, our neighborhood relations might become more harmonious. However, on the other hand, it's important to consider their preferences and follow medical advice. I'll also pay attention to Avery’s personal well-being!
Key Questions Delivery - Q7
Chatbot: Thank you so much for being with me for such a long time! Let's now discuss our final topic: if you were traveling or engaging in activities with Avery, would you have any feelings of frightened or being threatened regarding the potential implications of your interaction?
User: yes
Chabot: You would feel threatened or frightened about the potential implications of that interaction. Thank you for sharing your perspective.
Chatbot: May I have your reason as well?
User: If he cannot control himself and have a bad physical condition, it’s hard for us to travel together. I’m afraid that he cannot play well with me, he will feel uncomfortable, and me too.
Chatbot: I understand your concerns. It's natural to worry about both his well-being and your travel experience.
Chatbot: Hmmmm for me, if I were traveling or engaging in activities with Avery, I would likely have mixed feelings. On one hand, I would empathize with their struggles and try to offer support. On the other hand, I might also feel frightened about the challenges and uncertainties that could arise due to their unhappy and angry conditions.
Ending
Chatbot: That’s all for my questions. Thank you for doing this with me!
Chatbot: Now, do you have any questions? Feel free to let me know!
User: What is this study for? / no
Chatbot: This study aims to understand people's attitudes towards mental illness and how social stigma is attributed in such situations. Your insights are valuable in helping us gain a better understanding of this important issue. Is there anything else you'd like to know or share about your experience in this study? / Great, thank you for participating! To wrap up, can you please share your overall impression of this study's approach to understanding attitudes about mental illness using vignettes? Your feedback is valuable.
User: The chat is good!
Chatbot: Thank you for your feedback! I'm glad to hear that you found the chatbot helpful. Your input is greatly appreciated.
Chatbot: Congratulations, Cynthia, we finish our fantastic discussion! I sincerely wish you all the best, have a good day!
Based on our collected qualitative data, we extract causal interrelationships to construct a causal knowledge graph. Through visual analysis of this graph, we illustrate the entities and relationships in a format accessible to social scientists. Stay tuned for our upcoming preprint on causal-knowledge-graph construction methodology!
Dynamically examine stigma patterns
Compare individual stigma formations
Trace conversational contexts
Display causality details
Display psychological-construct details
Filter psychological constructs
Map construct-centered subgraphs
October 30, 2024: First version release! This initial release introduces our core visual analysis system.
December 15, 2024: Second version release will feature:
User-driven data input capability and custom theoretical framework integration
Semi-automatic theoretical framework extension based on built causal knowledge graphs
Please cite this article as:
@article{meng2024exploring,
title={Exploring the Potential of Human-LLM Synergy in Advancing Qualitative Analysis: A Case Study on Mental-Illness Stigma},
author={Meng, Han and Yang, Yitian and Li, Yunan and Lee, Jungup and Lee, Yi-Chieh},
journal={arXiv preprint arXiv:2405.05758},
year={2024}
}
Our visualization system's source code is publicly available on GitHub: https://github.com/guguqiui/visualization-system
We welcome contributions and feedback through GitHub issues!
We actively encourage collaboration and welcome feedback on this project!
For inquiries about research opportunities or other details, please contact:
Han Meng han.meng@u.nus.edu