Negative e-WOM and Customer Loyalty in Vietnam's Luxury Jewelry Sector: The Mediating Roles of Trust Crisis and Corporate Brand Image and the Moderating Role of Perceived Social Media Crisis Management Capability

Research design This study employed a cross-sectional, explanatory quantitative design to examine the proposed relationships among negative electronic word-of-mouth, perceived product authenticity risk, perceived price unfairness, service recovery failure, trust crisis, corporate brand image, perceived social media crisis management capability, and customer loyalty in Ho Chi Minh City’s luxury jewelry sector. The unit of analysis was the individual luxury jewelry customer. This context was selected because luxury jewelry purchases involve high value, information asymmetry, and strong reliance on certification, brand reputation, after-sales service, and trust. Customers may therefore face difficulty independently assessing authenticity, certification reliability, pricing fairness, and service commitments, making online information and firm responses important evaluative signals. Research method Data were collected using a structured self-administered questionnaire measuring respondents’ perceptions of negative online information, authenticity risk, price fairness, service recovery, trust crisis, corporate brand image, social media crisis management capability, and customer loyalty. All constructs were assessed using reflective multi-item scales on a five-point Likert scale from 1 = strongly disagree to 5 = strongly agree. The questionnaire comprised two sections. The first collected screening and demographic information, including gender, age, monthly income, luxury jewelry purchase experience, and exposure to online jewelry-related information. The second measured NEWOM, PAR, PPF, SRF, TC, CBI, SMCRC, and CL. To ensure consistent brand-level evaluations, respondents were instructed to select one luxury jewelry brand they had purchased from or seriously considered and answer all construct items with reference to that focal brand. Sample size, sampling procedure, and common method bias assessment The study employed non-probability purposive and convenience sampling to recruit luxury jewelry customers in Ho Chi Minh City who could meaningfully evaluate online information, product authenticity, pricing, after-sales service, brand image, and loyalty. Eligible respondents were at least 18 years old, had purchased luxury jewelry, and had been exposed to relevant online reviews, brand-related social media communication, or customer discussions. Data were collected online from January 16 to March 16, 2026, following ethical approval. The questionnaire was distributed through social media channels, customer networks, and brand-related discussion groups. Participation was voluntary and anonymous, and respondents were required to provide electronic informed consent before accessing the questionnaire. An a priori power analysis using G*Power 3.1 indicated a minimum sample size of 146, based on linear multiple regression with a medium effect size (f² = 0.15), α = 0.05, statistical power = 0.95, and six predictors in the trust crisis equation (Cohen, 2013; Faul et al., 2009). Of 500 distributed questionnaires, 453 valid responses were retained after excluding incomplete, ineligible, inconsistent, straight-lining, and duplicate cases.