Research summary
Different kinds of credibility work on different audiences — expertise sells to the committed fan, appeal sells to everyone else
Summarising the article
The Impact of VTubers and Streamers on the Purchase Intention of Otaku and Non-Otaku Respondents: A Comparative Study
Assoc. Prof. Dr. Smith Boonchutima · Basic and Applied Social Psychology · 2023
The question
Two kinds of gaming influencer — a VTuber, who is a virtual character, and a live streamer, who is a real person — affect viewers' purchase intentions differently. Which dimensions of credibility matter, and to whom?
How it was done
An online survey of 686 Thai female gamers who follow at least one female gaming influencer, mean age 24.79. Of these, 105 identified as Otaku and 581 did not. The survey measured three source-credibility dimensions — expertise, attractiveness and trustworthiness — using two real influencers as stimuli, one a VTuber and one a streamer, and analysed the data with mean comparisons and regression.
What it found
- With both influencer types combined, all three dimensions significantly predicted purchase intention — expertise (β = .330), trustworthiness (β = .237) and attractiveness (β = .188).
- Split by type, the picture changes sharply. For the VTuber, the predictors were expertise (β = .423) and trustworthiness (β = .258), while attractiveness was not significant (β = .108). The model explained 57.2% of the variance.
- For the streamer it reversed. The predictors were trustworthiness (β = .438) and attractiveness (β = .230), while expertise was not significant (β = .039). The model explained 46.7% of the variance.
- Otaku respondents rated expertise significantly higher (Cohen's d = 0.518, a medium effect) and reported higher purchase intention (d = 0.615, medium to large).
- The ranking of factors differed by audience. For Otaku, expertise mattered most (β = .333), then trustworthiness (.256) and attractiveness (.178). For non-Otaku, attractiveness had the strongest influence and trustworthiness was not a significant predictor.
Why it matters
This is directly usable when choosing a partner and writing a brief. If the audience is committed to the category, what has to be demonstrated is genuine knowledge of it, not image — so the content should show skill and real understanding of the product. For a general audience, familiarity and liking do more work. And because a VTuber is a designed character, the finding that attractiveness did not predict anything for that type suggests viewers judge the character by what it can do rather than by how it was drawn.
Limits to know before citing this
The sample is Thai female gamers only, and the authors note this limits transfer to other cultural contexts. The measures are self-reported intentions, not observed purchases. The Otaku group numbered only 105 against 581 in the other group, so the comparison rests on very unequal cells. Only two influencers served as stimuli, and the study did not separate content types. Purchase intention is not sales.
Cite this work
Pick a style and copy it straight into your bibliography
Boonchutima, S., & Surakanon, A. (2023). The impact of VTubers and streamers on the purchase intention of otaku and non-otaku respondents: A comparative study. Basic and Applied Social Psychology, 45(2-3), 63-79. https://doi.org/10.1080/01973533.2023.2208246
Generated from DOI registry metadata · check it against your institution's style guide before submitting
The original article is published under the CC BY-NC 4.0 licence; copyright remains with the authors and the publishing journal. The centre does not host a copy of the file; every link goes to the journal’s own repository, so readers always get the current version even if an erratum is issued later.
The summary on this page is written by the centre and is not text from the original article. For academic citation, cite the original article through its DOI.
Research with us
We welcome collaboration on research, programme evaluation, and joint publication.