Valuation of candidate brand equity dimensions and voting intention: alternative polling data in the Spanish presidential election

The study investigates whether and how candidate-based brand equity (CBBE) relates to voters’ intentions, framing political candidates as brands. Drawing on marketing and political branding literature, the authors posit that voters form perceptions across multiple brand equity dimensions—awareness, associations, perceived quality, loyalty, and emotion—that can influence voting intention. The research poses two main aims: (1) to determine if a multidimensional CBBE model provides unique brand profiles for different candidates (H1 and sub-hypotheses H1a–H1e), and (2) to test whether overall CBBE positively affects candidate-related voting intention (H2). The context is Spain’s shift from a long-standing two-party system to a multi-party system, where candidate branding may offer informational shortcuts and emotional connections that shape voter behavior.

The background synthesizes brand equity theory from Aaker (1991, 1996) and Keller (1993, 2001), where customer-based brand equity is a multidimensional construct encompassing awareness, associations, perceived quality, and loyalty (proprietary assets excluded here as an objective factor). Prior operationalizations include direct and indirect measures, with political adaptations leading to voter-based and candidate-based brand equity models distinguishing party (PBBE) and candidate (CBBE) components. The human brands perspective suggests candidates function as identity-laden brands distinct from their parties. Research shows generalized impressions of politicians (warmth, competence, associations, emotions) can shape voting intentions. The Spanish electoral system context (d’Hondt rule, 350-seat Congress, emergence of multi-party competition since 2015) underscores the relevance of candidate-focused brand evaluations. From this literature, the authors derive hypotheses: H1 that the CBBE model yields distinct brand profiles across candidates via its dimensions (H1a–H1e for each dimension’s influence), and H2 that higher CBBE leads to higher voting intention.

Design: Cross-sectional survey administered in Spanish across seven Spanish cities using non-probabilistic discretionary quota sampling. Participation was voluntary and uncompensated. A pre-test (n=56; ages 16–64) ensured scale adequacy and demographic balance; the final sample was N=375. Sample: 51.4% male; mean age 30.4 years (SD=12.4; range 18–81). Education skewed to higher education (72.4%). Party preferences covered major statewide parties. All participants reported awareness of all five candidates, a criterion for inclusion. Candidates rated: Mariano Rajoy (PP), Pedro Sánchez (PSOE), Albert Rivera (Ciudadanos), Pablo Iglesias (Podemos), Alberto Garzón (IU). Each respondent evaluated all five candidates on all items. Measures: 17 items total—14 for CBBE across five dimensions and 3 for voting intention (VI); 5-point Likert scales (1=strongly disagree; 5=strongly agree). CBBE dimensions: Awareness (3 items), Associations (3), Perceived Quality (3), Loyalty (3), Emotion (2). VI: 3 items reflecting willingness and likelihood to vote for the party associated with the candidate and trust in voting a party promoted by the candidate. Items adapted from Aaker (1991, 1996, 2011) and related scales; items with factor loadings <0.40 were removed, yielding the 17-item final model. Table 2 reports psychometrics: composite reliabilities ≥0.868, AVE ≥0.687, Cronbach’s alpha ≥0.751; RMSEA=0.055. Modeling approach: CBBE specified as a hierarchical construct—reflective first-order dimensions combined into a second-order formative construct (CBBE index). Assumption: first-order dimensions need not correlate (formative specification). Analysis employed PLS-SEM for prediction emphasis and accommodation of formative constructs. Sample size adequacy confirmed (N=375 > minimum ≈91 for medium effect, α=0.05, power=0.80). Measurement validation: factor loadings exceeded 0.71 on intended constructs; VIF values for formative indicators ranged 1.5–3.0 (acceptable). Internal consistency: composite reliability >0.87; alpha >0.70. Convergent validity: AVE >0.50; discriminant validity via Fornell–Larcker criterion. Structural modeling tested H2 (CBBE → VI). Importance–Performance Map Analysis (IPMA) was conducted per candidate to assess the practical leverage of each dimension. Data handling: For structural tests of H1/H2, candidate-specific responses were aggregated to participant-level mean scores representing general assessments; separate per-candidate descriptive profiles were also computed (Table 3).

• Structural path: Candidate brand equity showed a strong positive effect on voting intention (β=0.62), explaining approximately 39% of VI variance (R²≈0.39), supporting H2.
• Second-order weights/importance for CBBE indicators (overall): perceived quality (≈0.34) and loyalty (≈0.34) were the most influential, followed by emotion (≈0.22), awareness (≈0.22), and associations (≈0.20), supporting H1a–H1e in establishing unique profiles.
• Importance–Performance (overall):
  • High importance, low performance: loyalty (importance 0.344; performance 2.296), perceived quality (0.342; 2.671) → key levers for improvement.
  • Low importance, high performance: awareness (0.218; 4.053), associations (0.201; 3.961).
  • Low importance, low performance: emotion (0.222; 2.499).
• Candidate-level brand equity means: Rivera 3.42 (highest), Iglesias 3.18, Sánchez 2.77, Garzón 2.72, Rajoy 2.60. Candidate-related voting intention means: Rivera 2.83, Iglesias 2.43, Sánchez 2.22, Garzón 2.21, Rajoy 2.20.
• Candidate differences: Rajoy and Sánchez showed higher importance (coefficients) for loyalty, perceived quality, and emotion than the overall model but lower performance on these dimensions, indicating room for improvement. Garzón resembled Rajoy/Sánchez on awareness and associations but differed on other dimensions. Iglesias and Rivera showed lower importance and performance on loyalty, perceived quality, and emotion but higher awareness and association than overall averages.
• Psychometrics: All reflective constructs demonstrated strong reliability (composite reliability ≥0.868; alpha ≥0.751), convergent validity (AVE ≥0.687), and discriminant validity; formative collinearity acceptable (VIF 1.5–3.0). RMSEA=0.055 for the measurement model.

Findings indicate that multidimensional candidate-based brand equity captures voter-relevant perceptions that significantly predict voting intention, addressing the study’s core question. The strong path from CBBE to VI (β=0.62) demonstrates that aggregate perceptions—especially perceived candidate quality and loyalty—are pivotal in shaping intentions. IPMA shows these high-impact dimensions underperforming, highlighting strategic opportunities: enhancing perceived competence/quality signals and building attitudinal loyalty could most effectively elevate overall brand equity and, in turn, voting intentions. Awareness and associations are generally saturated across candidates, offering limited marginal gains except for specific cases (e.g., Garzón). Emotion currently contributes less but may represent an emergent lever as campaigns evolve. Candidate-specific profiles underscore that different candidates require distinct brand management priorities, validating the premise that CBBE yields actionable, individualized brand maps (H1). Practically, CBBE complements traditional polling by explaining the “why” behind voter choices and indicating which brand dimensions to manage to influence knowledge, feelings, and behavior.

The study advances political marketing by operationalizing and validating a hierarchical, multidimensional candidate brand equity model that predicts voting intention. It demonstrates that CBBE can serve as an actionable alternative to standard polling by both forecasting intentions and diagnosing which brand dimensions to manage. Empirically, perceived quality and loyalty are the most influential yet underperforming dimensions, suggesting strategic focus areas for campaign optimization. Newer party candidates with higher CBBE scored higher on associated voting intentions, aligning with observed electoral outcomes. Future research should test the model across countries and political levels, incorporate moderators such as party-based brand equity, refine voting intention measures, further develop the emotion dimension, and apply longitudinal designs to track and manage CBBE dynamics over time.

• Exploratory design with a non-probabilistic quota sample from selected cities may limit generalizability; sample exhibited bias toward Ciudadanos voters.
• Focus on candidate brand equity measures may omit other determinants of voting behavior; party brand equity not modeled as a moderator.
• Theoretical ambiguity regarding loyalty’s role (as a dimension vs. outcome) may affect construct specification.
• Cross-sectional data preclude causal inference; lack of longitudinal tracking limits insights into temporal dynamics.