Media Effects Research Lab - Research Archive
Effects of Conversational Partner, Message Interactivity, and Group Identity on Mental Health Disclosure Intention
Student Researcher(s)
Eunchae Jang (Ph.D Candidate);
Yansheng Liu (Ph.D Candidate);
DongWook Jeong (Ph.D Candidate);
Faculty Supervisor
This paper was based on a project as part of the “COMM506: Research Methods in Communications” course.
INTRODUCTION
Mental health problems became prevalent, which could cause severe self-injuries. It is known that disclosure has benefits; however, some people hesitate to disclose their mental health issues to others. This study investigated individuals’ willingness to disclose their mental health-related problems under different conditions, the identity of the conversational partner (AI/human), message interactivity (high/low), and group identity (in-group/out-group).
RESEARCH QUESTION / HYPOTHESES
H1: Participants in AI conditions will be more likely to disclose their mental health problems than those in human conditions.
H2: Participants in the high interactivity condition are more likely to disclose their mental health problems than those in the low-interactivity condition.
H3: The relationship stated in H1 would be pronounced in the high interactivity conditions than the low interactivity conditions.
RQ1: For individuals, controlling for AI usage, existing stress level, and US affiliation, what is the relationship between types of group identity of therapist (in-group vs. out-group) and the degree of disclosure intention?
RQ2: Is there an interaction effect between group identity and interactivity on self-disclosure intention?
RQ3: Is there an interaction effect between identity of conversational partner and group identity on disclosure intention?
RQ4: Is there a three-way interaction effect between identity of conversational partner, message interactivity, and group identity?
METHOD
This study conducted an online experiment with a 2 (conversational partner: AI vs. human) x 2 (interactivity: high vs. low) x 2 (group identity: ingroup vs. outgroup) between-subject designs. We recruited 174 US citizens as participants, and they were randomly assigned to one of the eight conditions. After watching the stimulus (see Figure 1[1]), participants were asked to complete a questionnaire that included the manipulation check questions, dependent variable, and control variables.
RESULTS
A three-way analysis of covariance(ANCOVA ) was run to examine the relationship between conversational partner, message interactivity, group identity, and disclosure intention when controlling AI usage, existing depression, and US affiliation. The results show that the main effect of message interactivity was marginally significant. Participants in the high interactivity condition showed a slightly higher disclosure intention than those in the low interactivity condition. In addition, this study ran the PROCESS macro (model 4) to examine the mediation role of perceived contingency on the relationship between message interactivity and disclosure intention. In other words, high message interactivity led to greater perceived contingency, which in turn increased the disclosure intention.
CONCLUSIONS/DISCUSSION
This study investigates how the identity of conversation partners, message interactivity, and group identity affect disclosure intention in the context of mental health. Even though the current study could not find any significant effects except interactivity, the direction of effects was consistent with the proposed hypotheses. First, the mean values of the AI condition were higher than the human condition. This supports the previous findings that facilitating AI therapists can help reduce people’s mental health problems; however, this also aligns with the findings from other research that showed no differential effect of the conversational partner, supporting the Computer as Social Actors paradigm that people apply the same social rules when interacting with computers. Second, although interactivity was marginally significant, it can be said that users could be more likely to disclose when they feel that the therapist is contingently responding to their words while responding quickly. Lastly, one possible reason for the nonsignificant effect of group identity could be due to the severity level. Mild symptoms of depression might not have been enough for participants to consider their negative evaluations from others or stigmatization. However, this study is not free from limitations, such as stimuli, the severity level of mental health problems, and the manipulation of group identity as nationality. Furthermore, it is recommended for future studies to conduct longitudinal studies to show more relational outcomes with conversational partners because it is known that people build intimacy by increasing the degree of self-disclosure over time.
For more details regarding the study contact
Dr. S. Shyam Sundar by e-mail at sss12@psu.edu or by telephone at (814) 865-2173