Emotions and customer experience in US retail banking: a PANAS-based structural equation model

  • 5 Views
  • 1 Downloads

Creative Commons License DMCA.com Protection Status
This work is licensed under a Creative Commons Attribution 4.0 International License

Type of the article: Empirical Research Article

Abstract
Affective responses are an important component of customer experience, but evidence on how positive and negative affect relate to downstream outcomes in US retail branch banking remains limited. This study examines generalized affect associated with branch interactions, measured with the Positive and Negative Affect Schedule (PANAS), and its associations with satisfaction, loyalty, and willingness to recommend. Cross-sectional survey data were collected in April 2024 from 400 US retail banking customers through the Pollfish online panel and analyzed using confirmatory factor analysis and covariance-based structural equation modeling. Standardized structural estimates show that Positive Affect is positively associated with Satisfaction (β = 0.441, p < 0.001), whereas Negative Affect is negatively associated with Satisfaction (β = –0.292, p < 0.001). Satisfaction is associated with Loyalty (β = 0.504, p < 0.001) and Recommendation (β = 0.248, p < 0.001), while Loyalty is also associated with Recommendation (β = 0.572, p < 0.001). The model explains 28.0% of the variance in Satisfaction, 25.4% in Loyalty, and 53.2% in Recommendation. Overall fit is satisfactory (χ2(400) = 545.123, CFI = 0.991, TLI = 0.990, RMSEA = 0.030, 90% CI [0.023, 0.036]). PANAS is applied as an established affect measure rather than newly validated. Because no specific recent branch encounter or recall period was defined, affect scores are interpreted as generalized recalled affect, and the cross-sectional design precludes causal inference.

view full abstract hide full abstract
    • Figure 1. Proposed structural equation model
    • Table 1. Realized sociodemographic distribution of the sample
    • Table 2. Latent variables considered in the model
    • Table 3. Rechecked AMOS regression weights and standardized estimates
    • Table 4. Reliability and convergent validity
    • Table 5. Discriminant validity of the measurement model (HTMT)
    • Table 6. Overall structural SEM fit indices
    • Conceptualization
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Data curation
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Formal Analysis
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Investigation
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Methodology
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Project administration
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Resources
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Software
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Supervision
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Validation
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Visualization
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Writing – original draft
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín
    • Writing – review & editing
      Carlos Alberto Espinosa Fernández, Ainhoa Rodríguez Oromendía, Iñigo Tejera Martín