Dissertation
Support in the context of human-machine communication: a test of competing perspectives
University of Iowa
Doctor of Philosophy (PhD), University of Iowa
Spring 2023
DOI: 10.25820/etd.007001
Abstract
Support in the context of human-machine communication (SHMC) is the human use of machine agents in largely informal ways to intentionally seek or receive resources to buffer stress. However, researchers traditionally consider supportive communication as an interpersonal exchange between humans. In contrast, some recent research contends that machine agents pose a viable option to aid loneliness and depression and address human social support needs. I refer to these frameworks as the impairment and improvement perspectives, respectively. Whereas the impairment perspective suggests that machine agents will negatively influence support, the improvement perspective asserts their influence will be positive. Because previous research in both support and HMC suggest competing predictions for how machines will affect processes of supportive communication, this project seeks to synthesize and build upon understandings of how perceptions of machines influence processes and outcomes of supportive interactions by examining the role of seeking and processing supportive messages between humans and robot conversational partners.
In a 2X2 (human and chatbot support partners using high or low person-centered messages) between-groups experimental design, I found that participants were more likely to seek emotional support from the chatbot than they were from human partners and were also more likely to report direct support seeking behaviors with the chatbot rather than the human partners. I then found that participants reported no significant differences in perceptions of uncertainty or social presence from the chatbot than with humans, but reported lower levels of liking for the chatbot. These impressions, however, did not contribute to indirect effects between provider status and support seeking strategy (direct or indirect). I finally explored the role of verbal person-centeredness (VPC, Burleson, 1987) and how people perceive machine agents on several interpersonal and supportive communication outcomes. Consistent with previous research, a direct effect was found for message strategy such that high person-centered messages were evaluated more highly than low person-centered messages on all dependent support evaluation and outcome variables. Additionally, an interaction effect was found between bots and humans using low and high person-centered messages. Results extend the literature on support seeking, provision, and the quality of these processes in the context of HMC.
Details
- Title: Subtitle
- Support in the context of human-machine communication: a test of competing perspectives
- Creators
- Austin J. Beattie
- Contributors
- Kate Magsamen-Conrad (Advisor)Andrew C High (Advisor)Autumn P Edwards (Committee Member)Rachel McLaren (Committee Member)Kembrew McLeod (Committee Member)Rachel Young (Committee Member)
- Resource Type
- Dissertation
- Degree Awarded
- Doctor of Philosophy (PhD), University of Iowa
- Degree in
- Communication Studies
- Date degree season
- Spring 2023
- Publisher
- University of Iowa
- DOI
- 10.25820/etd.007001
- Number of pages
- xiii, 166 pages
- Copyright
- Copyright 2023 Austin J. Beattie
- Language
- English
- Date submitted
- 04/25/2023
- Date approved
- 06/07/2023
- Description illustrations
- illustrations, tables, graphs
- Description bibliographic
- Includes bibliographical references (pages 125-142).
- Public Abstract (ETD)
- Support in the context of human-machine communication (SHMC) is the human use of machine agents like chatbots and AI to help with stress. Some researchers consider the process of supportive communication to be one that occurs only (or should only occur) between 'humans'. In contrast, some recent research argues that machine agents might possess some ability to help. I refer to these frameworks as the 'impairment' and 'improvement' perspectives, respectively. Whereas the 'impairment' perspective suggests that machine agents like chatbots and AI will hurt the ways people seek and receive help, the 'improvement' perspective argues machine agents can help. Because previous research has found support for both perspectives, this project seeks to synthesize them by examining the role of looking for help, and evaluating help, between humans and machine agent conversational partners. In this study I found that participants were more likely to seek emotional support from chatbots than they were from humans and were also more likely to report direct support seeking behaviors chatbots than the human partners. I then found that participants did not report feeling more uncertain or to the chatbot than with humans, but that they reported lower levels of liking for the chatbot. These feelings, however, did not change the strategies people reported using asking for help. I finally explored the role of how closely support providers (human or machine agent) adapt their messages to the feelings of others on how people evaluated the quality of the help they received. Consistent with previous research, messages that adapted more to the individual stressors of their recipients were associated with better outcomes than messages that adapted less to their targets. Results extend the literature on support seeking, provision, and the quality of these processes in the context of HMC.
- Academic Unit
- Communication Studies
- Record Identifier
- 9984428938702771
Metrics
6 File views/ downloads
23 Record Views