Skip to content

Amy's Lab Notebook

Jump to bottom
A.X edited this page Jul 4, 2023 · 18 revisions

[June 5 - June 18]

Meeting Notes

To Do lists

  • read lit review (sent in slack) >> how do you measure things like closeness or empathy
  • review live typing set-up
  • Look into semantic search similarity

Task Updates

Lit review:

  • Effects of Self-Consciousness and Social Anxiety on Self-Disclosure Among Unacquainted Individuals

"Public self-consciousness does not seem to be related to the level at which subjects disclose but rather appears to influence subjects' beliefs about how their partners view them." "It appears that the different subscales of the self-consciousness scale may influence the acquaintance process in rather divergent manners"

  • Can anthropological techniques apply in this case? If certain demographic types may influence interaction patterns?
  • "perception of disclosure" may be an interesting metric to keep track of in long-term communication -does it change it downstream disclosure or personal relationships?
  • "I Hear You, I Feel You"

"We can infer that the specifics of social interaction between a chatbot and its users can affect self-disclosure length. Additionally, length of use (as measured by experiment day) and chatbot variant both might influence the participants’ willingness to disclose their thoughts and feelings to a chatbot.” "However, despite all three groups being told that the chatbot represented a counselor from their local area, sharp inter-group differences emerged in how they perceived its persona" "But I don’t have to care about the chatbot. I can just talk about myself and focus on how I feel." “This implies that the users may assume the chatbot has more intelligence than it actually does, which might lead to users not reaching out to professionals for proper help.”

  • How does they presence of researchers affect the sense of anonymity and, in turn, the willingness of participants to self-disclose
  • Bot disclosure level seemed to have the biggest effects on feelings discussed
  • Are the 3 groups having an effect both in “presenting instructions for how this chat is supposed to go” and “what I am willing to share?”
  • How did this study measure trust, enjoyment and intimacy?
  • Why did trust not increase over time?
  • “Together but not together”

"Messaging was seen as a way to “uphold an image where there can be no room for error, vulnerability, and a close human connection”. The applicability of live-typing for interpersonal communication was seen as a threat since it limited participants’ ability for selective self-presentation as they could no longer edit their messages in a “controlled and socially desirable fashion”

  • The diverging experiences with a potential live typing feature reminds me of the togglable "read" function in messages. Also in Snapchat, they would play the typing sound and indicate a chat had been started even if you did not send anything.
  • How does the role of live typing play in facilitating personal conversations where one or both parties are preoccupied with other tasks? How do these set of typing indicators compare to the current paradigm of three bubbles, stop, pop up? What is the comparative richness within the context of user patters
  • Enhancing user experience with the conversational agent for movie recommendations

"1) How does self-disclosure between a user and a CA affect user satisfaction and intention to use? 2) How does reciprocity between a user and a CA affect user satisfaction and intention to use?" "Taylor and Hinds, 1985).Berschied et al. (1989) provided the basis for the intimacy measurement items that are used to measure user perception of closeness and familiarity after conversing with an agent. Trust items consisted of measures of the reliability, honesty, and trustworthiness of the agent " "However, using a real artificial 3D conversation agent, or using virtual reality, might generate a higher social presence"

  • Self-disclosure and intimacy did not seem to be directly linked
  • Trust was strong, but what did trust mean in this case? Trust in competency?
  • Is it unerving to have a CA in a setting like netflix?
  • Used a 2D avatar -> what is social presence
  • Apendix A for survey items

General Thoughts:

  • Online ethnography of media richness in high self-conscious individuals → lex and grinder etc? Characters? What moves people into online spaces?

→ create anonymous chat groups → do certain patterns in self-disclosure emerge? Wb is a mixed group of CA agents and high anxious individuals

→ Computed-mediated group conversations

→ How does mediums richness affect saliency?

  • Medium richness → longitudinal study?

  • Is there a long-term divergence in the way people treat CA, does that then inform the way they perceive and interact with humans?

To-read:

  • Irwin Altman and Dalmas A Taylor. 1973. Social penetration: The development of interpersonal relationships. Holt, Rinehart & Winston.

  • Ellen Berscheid, Mark Snyder, and Allen M Omoto. 1989. The relationship closeness inventory: Assessing the closeness of interpersonal relationships. Journal of Personality and Social Psychology 57, 5 (1989), 792.

  • Yi-Chia Wang, Moira Burke, and Robert Kraut. 2016. Modeling Self-Disclosure in Social Networking Sites. In Proceedings of the 19th ACM Conference on Computer-Supported Cooperative Work & Social Computing (CSCW ’16). ACM, New York, NY, USA,

  • Mark, G., Becker, B., 1999. Constructing social systems through computer-mediated communication. Virtual Real. 4 (60). http://dx.doi.org/10.1007/BF01434995.

  • Robin L Wakefield, Kirk L Wakefield, Julie Baker, and Liz C Wang. 2011. How website socialness leads to website use. European Journal of Information Systems 20, 1 (2011), 118–132.

  • Richard L Daft and Robert H Lengel. 1986. Organizational information requirements, media richness and structural design. Management science 32, 5 (1986), 554–571

[ June 18 - June 30 ]

Action Items

Task Updates

Chat Repositories

Found this repo: https://github.com/kingofthestack/react-chat-window/tree/master/demo, which is a straightforward react-based chat app. I downloaded the demo and then tried to integrate the live-typing features into it by importing live-typing lib into the node_modules folder. Then I used the function in the launch components, putting the response element in the DOM. I then encountered a "reserved keyword error" which indicated I was running into issues parsing the typescript in live typing. I tried several different solutions, primarily centering on adding/modifing the config/eslint files. I think it may also be due to the directory structure, so I will also have to look more deeply into that. At this point, it had been around 2 hours so I decided to wait until I could touch base with someone more familiar with the structure. I have yet to fix the error, but it is on my to-do for next week.

"Hyperpersonal dimensions of technology, language, and cognition"

“As receivers, CMC users idealize partners based on the circumstances or message elements that suggest minimal similarity or desirability. As senders, CMC users selectively self-present, revealing attitudes and aspects of the self in a controlled and socially desirable fashion.”

“empirical tests have shown how CMC leads to more extreme impressions than FtF (Hancock & Dunham, 2001) and more positive relations over time compared to FtF”

“There is much less ‘‘leakage’’ in CMC since there is no unwanted nonverbal indication of undesirable affect or attitude”

“Editing behavior was also coded, using the following four definitions: (1) forward deletions; (2) destructive backspaces; (3) insertions (putting in spaces, letters, or punctuation where there previously had been none); and (4) replacements, such as changing capitalization to lower case, or changing words or letters. Three coders sampled 10 per cent of the tapes in common, with inter-rater reliabilities achieving Scott’s pi = .63, which was acceptable under these conditions.2”

  • Surprising that CMC leads to more positive reactions over time

  • I think these observations may vary greatly by age group and ‘texting culture’

  • I’m interested in if the authors validated their definition of a ‘high reward sources’ / had a metric that tracked this. ** They had pre-testing that showed their strongest preference was an opposite sex student. (maybe its a comfort thing)

  • Tracked editing behavior through tracking forward deletions, destruction backspaces, insertions, and replacements

  • More personal language with desirable targets

  • Mirrored language to accommodate who they were speaking to

  • More editing more mindfulness

"Modeling Self-Disclosure in Social Networking Sites"

We showed that women self-disclose more than men, and people who more strongly desire to manage the impressions they make on others self-disclose less. We then demonstrated that social network size was negatively associated with selfdisclosure, while network density and average tie strength had positive correlations with self-disclosure. Most of the results are consistent with those found or suggested by prior literature, which validate the effectiveness of the machine learning model we proposed.”

“As a sensitivity test, we replicated the analyses reported here on de-identified, aggregated posts from Facebook users in Australia and Singapore and discovered similar results.”

  • This study primarily focuses on text communication, but in SNs they often are heavily image focused -I wonder if the influence of images as a communication medium would change some of these observed patterns

  • Used humans to manually name model-produced directory

  • H4 is “Average tie strength will be positively correlated with self-disclosure” (pdf) , does that flip beyond a certain point toward anonymity?

"How website socialness leads to website use"

“The Technology Acceptance Model (TAM; Davis, 1989), as a by-product of the Theory of Reasoned Action (TRA) (Fishbein & Ajzen, 1975), has been a prominent starting point in the IS literature for examining the formation of attitude and behavioral intentions toward IT systems. Perceived usefulness (PU), perceived ease of use (PEOU) and enjoyment (Davis, 1989) determine attitudes, which in turn influence behavioral intentions to use an IT system.”

“An individual motivated to visit a website for what is primarily a utilitarian product (e.g., home improvement products or supplies) may respond differently to social cues compared to another who visits a website for a hedonic product”

  • I feel like the use of superimposed videos may have too strong unwanted implications for ux, in turn modifying the “sociability” of the site. Why not modify the copy instead?

  • How much of these findings are due to the stage of technology maturity at the time?

  • I think retail live chat/chatbots may be an extension of their interactive guides

Clone this wiki locally