Answers are grounded in this paper and cite their sources.
Digital brand engagement scores can obscure the participation they summarise. This study examines consumers’ online brand-related activities (COBRAs) using 17 indicators from 384 technology-start-up users in Guayaquil, Ecuador. Five graded response models, paired ordinal contrasts, conditional information, exact-score comparisons and firm-held-out payment models evaluate structure, ordering and scoring consequences. Three correlated domains outperformed one dimension (BIC reduction = 650.43) and every pairwise merger; correlations ranged from 0.575 to 0.710. Consumption, contribution and creation followed the expected order at all four empirical response cutoffs, preserving that aggregate order under every common increasing category recoding. All eight simultaneous participant-bootstrap contrasts were positive; seven remained conclusive in the nine-firm cluster sensitivity analysis. Only 33.6% of users had strictly descending domain scores. Exact overall-score groups containing descending and inverted profiles covered 91.4% of users, with 24.1–29.5% of domain-score variability unresolved within score groups. Separate domains yielded a payment-classification AUC of 0.614 versus 0.606 for the overall score, with nearly identical held-out log loss. The contribution specifies the hierarchy’s aggregate meaning and quantifies the behavioural detail obscured by scoring. Domain profiles support participation diagnosis; their additional commercial classification value requires separate validation. Cross-sectional evidence cannot identify individual transitions or causal commercial effects.
No takes yet. Share an insight, caveat, or question.
Hernández et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: