Exploring the impact of beauty vloggers' credible attributes, parasocial interaction, and trust on consumer purchase intention in influencer marketing

The study examines how beauty vloggers influence consumer purchase intentions for cosmetics through their source credibility attributes (trustworthiness, expertise, attractiveness), parasocial interactions (PSI), and the mediating role of trust. Motivated by the rapid rise of influencer marketing and the particular salience of beauty content on visually oriented platforms, the research addresses gaps in prior work that often assessed credibility dimensions in isolation and rarely considered PSI and credibility together, especially in emerging markets like India. The study draws on Ohanian’s (1991) source credibility model and Horton and Wohl’s (1956) parasocial interaction theory and poses three research questions: (RQ1) Do credible attributes of beauty vloggers help change consumers’ attitudes toward endorsed products? (RQ2) Does trust affect the relationship between source credibility attributes and purchase intention? (RQ3) Do repeated parasocial interactions by beauty vloggers affect purchase intention? It proposes a conceptual framework where source credibility (second-order construct with trustworthiness, expertise, and attractiveness as lower-order constructs) and PSI predict purchase intention directly, while trust mediates the effect of credibility on purchase intention. The study emphasizes the practical importance of understanding these mechanisms for cosmetic marketers seeking cost-effective strategies and highlights the role of trust-building and repeated interactions in shaping consumer behavior in India.

The literature review integrates source credibility theory and parasocial interaction theory to build hypotheses about influencer marketing effectiveness in the beauty domain. - Source credibility theory (Ohanian, 1991; Hovland & Weiss, 1951; McGuire, 1985) posits that endorser trustworthiness, expertise, and attractiveness drive message acceptance and persuasive impact. In cosmetics, both physical and social attractiveness can shape product perceptions. Prior influencer studies often assessed only parts of credibility (e.g., attractiveness alone or expertise alone), leading to mixed or incomplete findings. The review positions credibility as a multidimensional, second-order construct relevant to micro-celebrities/beauty vloggers. - Trust is conceptualized as a recipient’s confidence in the communicator (Morgan & Hunt, 1994; Mayer et al., 1995). Frequent interactions and perceived integrity, expertise, and reliability foster trust, which in turn drives positive attitudes and purchase intentions. Prior research suggests trust as a key mechanism linking influencer attributes to consumer behavior and as a mediator in related contexts. - Mediating role of trust: Consumers are more likely to adopt recommendations from influencers they trust; authenticity, transparency, and reliability build that trust. Trustworthiness and trust are related yet distinct; perceived trustworthiness contributes to the development of trust. Prior studies show trust can mediate effects from influencer characteristics to purchase intentions. - Parasocial interaction (PSI) theory (Horton & Wohl, 1956) describes illusionary, friendship-like relationships between audiences and media figures. With social media, PSI can be reciprocal (e.g., comments, Q&A, live sessions). PSI fosters familiarity and closeness that can translate into behavioral outcomes, including purchase intention. Hypotheses: H1: Source credibility of beauty vloggers positively influences trust toward beauty vloggers. H2: Source credibility positively influences consumers’ purchase intention. H3: Trust positively influences purchase intention. H4: Trust mediates the relationship between source credibility and purchase intention. H5: Parasocial interaction positively influences purchase intention.

Design: Quantitative, cross-sectional survey analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM) in SmartPLS 4. The model treats source credibility as a second-order reflective construct with three reflective lower-order dimensions: trustworthiness, expertise, and attractiveness. Measurement: Multi-item scales adapted from prior studies. Trustworthiness items from Ohanian (1991) and Teng et al. (2020); expertise and attractiveness items from Ohanian (1991), Ki & Kim (2019), and Teng et al. (2020); trust from Meng & Wei (2020); parasocial interaction from Lee & Watkins (2016) and Lou & Kim (2019); purchase intention from Jiménez-Castillo & Sánchez-Fernández (2019), Lou & Kim (2019), and Meng & Wei (2020). All items measured on 5-point Likert scales (1 = strongly disagree to 5 = strongly agree). Items with low loadings were removed to improve the measurement model (e.g., TW1; ATT7; PSI items PARA INT3, 6, 8, 9). Sample and data collection: Non-probability convenience sampling due to an unknown social media user population. Online survey administered to consumers aged 16–45 who regularly view beauty vloggers’ content (1–3 hours daily). Data collection: June 2022–March 2023 across North Indian regions (Punjab, Haryana, Chandigarh, Delhi NCR). Of 500 distributed questionnaires, 357 valid responses were retained after screening for inclusion criteria (response rate 71.4%). Ethical approval was obtained; informed consent was collected; for minor respondents (n=11), guardian consent was obtained. Analysis procedure: Two-stage evaluation using the repeated indicator approach. First, assessment of lower-order reflective constructs: indicator reliability (loadings target ≥ 0.708), internal consistency (Cronbach’s alpha, rho_A, composite reliability ≥ 0.70), convergent validity (AVE ≥ 0.50), discriminant validity via HTMT (≤ 0.85). Second, higher-order construct assessment for source credibility using latent variable scores; outer loadings of lower-order dimensions as indicators of the higher-order construct, with reliability and convergent validity achieved. Structural model: Checked collinearity (inner VIF < 3), estimated path coefficients with bootstrapping, tested mediation (trust), included gender as a control (non-significant), and assessed model fit (SRMR), explanatory power (R2), effect size (f2), and predictive relevance (Q2 predict; PLSpredict with RMSE comparisons versus linear models).

All hypothesized relationships were supported (PLS-SEM results): - H1: Source credibility → Trust: β = 0.628, t = 16.268, p < 0.001. - H2: Source credibility → Purchase intention: β = 0.226, t = 3.280, p = 0.001. - H3: Trust → Purchase intention: β = 0.416, t = 7.352, p < 0.001. - H5: Parasocial interaction → Purchase intention: β = 0.179, t = 3.326, p = 0.001. - H4 (mediation): Source credibility → Trust → Purchase intention (indirect effect): β = 0.261, t = 6.580, p < 0.001, indicating complementary partial mediation (direct path remains significant; indirect strengthens total effect). Model performance: - R2: Trust = 0.394; Purchase intention = 0.490. Adjusted R2: Trust = 0.392; Purchase intention = 0.486. SRMR = 0.069. - Effect sizes (f2): Source credibility on Purchase intention = 0.049 (weak); Source credibility on Trust = 0.650 (strong); Trust on Purchase intention = 0.202 (medium); Parasocial interaction on Purchase intention = 0.040 (weak). - Predictive relevance: Q2 predict for Trust = 0.385; for Purchase intention = 0.372; all indicator-level Q2 values > 0. PLSpredict RMSE values for purchase intention indicators were lower than linear model benchmarks, indicating superior out-of-sample predictive performance. Additional notes: - Gender control was non-significant for direct (β = −0.068, p = 0.392) and mediated (β = −0.095, p = 0.253) effects. Interpretation: - Credibility attributes and PSI both increase purchase intention; credibility shows a slightly stronger impact than PSI. Trust is a pivotal mechanism linking credibility to behavioral outcomes in the Indian beauty context.

The results demonstrate that beauty vloggers shape consumers’ purchase intentions when perceived as credible information sources and when audiences trust them. By validating the holistic source credibility framework, the study shows that trustworthiness and expertise are particularly impactful in driving trust and purchase intentions, with attractiveness also relevant in the beauty category. Trust partially mediates the link between credibility and purchase intention, underscoring trust-building as central to influencer effectiveness. Parasocial interactions also directly increase purchase intention, reflecting the role of repeated, quasi-friendship interactions in forging closeness and receptivity. The influence of credibility was found slightly stronger than PSI overall, aligning with the idea that in contexts like India, credibility and trust lay the groundwork for subsequent relational bonds. The findings address the research questions by confirming that credible attributes affect attitudes and purchases (RQ1), trust enhances the credibility–purchase intention relationship (RQ2), and repeated PSI positively affects purchase intention (RQ3). These insights highlight strategic levers for cosmetic marketers: prioritize credible, expert, and trustworthy influencers, while fostering ongoing interactions that cultivate closeness and ritualized viewing to strengthen behavioral responses.

This study advances influencer marketing research by jointly applying source credibility and parasocial interaction theories to explain consumers’ purchase intentions in the beauty vlogging context. Modeling source credibility as a second-order construct and establishing trust as a complementary partial mediator provides a more comprehensive account of how influencer attributes convert into behavioral outcomes. Empirically, both credibility and PSI significantly increase purchase intention, with credibility exerting a slightly stronger effect; trust is central to this mechanism. For practice, brands should select influencers high in trustworthiness and expertise, nurture audience trust, and encourage frequent, interactive content to build parasocial bonds. Future research can extend this framework across vlogging domains, explore audience-level moderators (e.g., self-concept, influencer–consumer congruence), examine regional/cultural variations, and assess virtual influencers using the same credible attributes model.

- Sample composition skews toward ages 18–30, with fewer respondents aged 31–45+, which may limit age-based generalizability. - Geographic scope is limited to North India (Punjab, Haryana, Chandigarh, Delhi NCR); results may differ across Southern, Eastern, and Western India. - The model focuses on two antecedent facets (source credibility and parasocial interactions) and excludes audience characteristics (e.g., self-concept, influencer–consumer congruence) and other potential drivers. - Domain focus is on beauty vloggers; findings may not fully generalize to other influencer categories (e.g., travel, gaming, health) without further testing. - Among credibility attributes, attractiveness remains less deeply explored; future studies could probe its nuances and boundary conditions. - Potential moderators (e.g., relationship strength) were not modeled; examining such moderators could enrich understanding. - As influencer marketing evolves with AI and virtual influencers, comparative analyses between human and virtual influencers are warranted.