Media Effects Research Lab - Research Archive
@VirtualInfluencer – Modality and Influencer Credibility of Virtual Influencers
Student Researcher(s)
Alice Nguyen (Ph.D Candidate);
Matt Checchio (Ph.D Candidate);
Faculty Supervisor
INTRODUCTION
Artificial intelligence (AI) has experienced rapid advancement in recent years giving rise to a new category of influencers known as Virtual Influencers. Our study examines virtual social media influencers and the difference in modality between image and video when it comes to influencer credibility. We theorize that an image followed by a video post of a virtual influencer will increase influencer credibility as opposed to just seeing an image. Furthermore, a parallel mediation of perceived realism and eeriness will mediate the relationship along with anthropomorphism as a moderator.
RESEARCH QUESTION / HYPOTHESES
For adults over 18 viewing Instagram posts featuring virtual influencers, controlling for social media use, what is the relationship between modality, anthropomorphism of influencer, and influencer credibility?
H1: An image followed by a video post of a virtual influencer will increase influencer credibility as opposed to just seeing an image.
H2a: Perceived realism will mediate the relationship between modality and influencer credibility.
H2b: Eeriness will mediate the relationship between modality and influencer credibility.
H3: Anthropomorphism moderates the relationship between modality and influencer credibility.
METHOD
We conducted a 2 (modality: image vs. picture) x 2 (anthropomorphism: high vs. low) mixed factorial design. Anthropomorphism was manipulated between-subject, and modality was manipulated within-subject. The experimental design included a mixture of scales to measure variables in credibility, perceived realism, eeriness, and a control for habitual social media usage. For hypothesis 1 and 3, we employed repeated measures MANOVA and for hypothesis 2 used PROCESS Macro for mediation effects (Model 4).
RESULTS
For H1, we found modality has a significant effect on credibility. For H3, the interaction effect between modality and anthropomorphism is significant. Perceived realism fully mediated the relationship between modality and influencer credibility, supporting H2a, but perceived eeriness did not mediate the relationship between modality and influencer credibility.
CONCLUSIONS/DISCUSSION
Our results support the ideas of virtual influencers and credibility between image and video modalities. When viewing an image of a virtual influencer, there is no difference in credibility between high and low anthropomorphism. However, for video, credibility increases for high anthropomorphism. Therefore, for practical implications, high anthropomorphic influencers should use both image and video to increase credibility in the marketing context. Finally, our study was limited with the internal validity of study design to pair stimuli accordingly and only using female virtual influencers.
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