Beyond emotions: Social cognitive predictors of COVID-19 vaccination intentions before and after vaccine roll-out

Context and Objectives

The COVID-19 pandemic has posed significant challenges for global public health, including the need to understand factors influencing vaccination intentions. Although considerable attention has been given to emotional aspects of people's responses to COVID-19, a notable gap exists in the literature regarding how socio-cognitive and emotional predictors interact to influence vaccination intentions across different cultural contexts and critical moments of vaccine rollout.

The study by Manoli et al. (2026) addresses this gap through a prospective longitudinal multinational design that examines how emotional and socio-cognitive factors influence COVID-19 vaccination intentions during two critical phases: before vaccine rollout (April-June 2020) and after (January-February 2021). This approach is particularly relevant because it allows examination not only of which factors predict vaccination intention, but also how these factors change and evolve in relation to significant contextual events such as vaccine rollout.

The primary objective of the study was to identify which socio-cognitive and emotional variables are robust predictors of COVID-19 vaccination intentions, and whether these relationships remain stable or change between the two evaluation periods. The authors also sought to determine whether psychological variables (perceived stress, psychological flexibility, positive affect, and prosociality) contribute significantly to predicting vaccination intentions, beyond the Health Belief Model (HBM).

Method

Participants

The study sample consisted of 586 adults (16.9% male) aged 18-79 years (M = 39.32, SD = 13.43), selected from a larger initial dataset of 9,565 respondents. The sample was recruited in the context of a large multinational project that included data from 33 countries. The predominantly female composition of the sample (83.1%) is a recognized limitation by the authors that should be considered when generalizing results.

Design

The study employed a prospective longitudinal design with two clearly defined evaluation points. The first evaluation timepoint (Time 1) took place during April-June 2020, before COVID-19 vaccine rollout. The second evaluation timepoint (Time 2) occurred during January-February 2021, after vaccine rollout began in many countries. This design enables capture of changes in key variables and examination of how predictors of vaccination intention might vary according to temporal context.

Data were collected via an online survey administered through RedCap, available in 18 languages, which facilitated multinational and multilingual participation.

Intervention/Conditions

No specific intervention is described in this study. Rather, the study is observational and leverages natural events (vaccine rollout) that occurred between the two evaluation points as a "natural experiment" to examine how relationships between variables change.

Instruments

The authors employed a comprehensive set of validated measures:

1. Perceived Susceptibility and Severity (based on the Health Belief Model): Assessed with 3 items each on a 5-point scale (α = 0.86). These measures evaluate how susceptible respondents feel to COVID-19 and how severe they believe the illness would be.

2. COVID-19 Risk Self-Efficacy: An adapted version of the New General Self-Efficacy Scale was used, comprising 5 items (α = 0.88). This scale measures participants' confidence in their ability to protect themselves against COVID-19.

3. Perceived Stress Scale: The 10-item version was administered (α = 0.92), which evaluates the degree to which participants perceive their lives as unpredictable, uncontrollable, and overloaded.

4. PsyFlex: A 9-item measure (α = 0.91) was used to assess psychological flexibility, representing the capacity to maintain contact with the present moment while pursuing personal values, even in the presence of difficult thoughts and emotions.

5. Expanded PANAS: An expanded version of the Positive and Negative Affect Schedule was employed on a 7-point scale (positive affect α = 0.95) to measure affective states.

6. Prosocialness Scale for Adults: A 6-item measure (α = 0.90) was used to assess prosocial tendencies.

7. Vaccination Intention: Measured only at Time 2 using a 7-point Likert scale, assessing participants' willingness to be vaccinated against COVID-19.

Analysis

Statistical analysis was rigorous and multifaceted:

• Kruskal-Wallis tests were used to examine differences in sociodemographic variables.
• Paired t-tests were employed to evaluate changes in psychosocial variables between Time 1 and Time 2, allowing examination of how these variables evolved in relation to vaccine rollout.
• Cumulative link models (CLMs) were specified using R's ordinal package. Four different models were tested, presumably to examine different combinations of predictors and their ability to predict vaccination intentions.

Ethical approval was obtained from Cyprus National Bioethics Committee.

Results

Changes in Psychosocial Variables between Time 1 and Time 2

The study documented significant changes in several key psychosocial variables between the pre-vaccine rollout period and the post-rollout period:

• Self-efficacy: There was a significant decline in COVID-19 risk self-efficacy from Time 1 to Time 2 (t(595) = 4.65, p < .001, d = .19), suggesting that people's confidence in their ability to protect themselves against COVID-19 decreased during this period.

• Prosociality: Prosociality also decreased significantly (t(595) = 6.79, p < .001, d = .28), which could reflect shifts in social orientations during the pandemic.

• Psychological flexibility: A significant though more modest decline was observed in psychological flexibility (t(595) = 2.84, p < .01, d = .11).

• Positive affect: Positive affect decreased significantly (t(595) = 2.46, p = .014, d = .10), consistent with previous studies documenting the emotional impact of the pandemic.

Predictors of Vaccination Intention

The main findings revealed interesting patterns in vaccination intention predictors:

Self-Efficacy: Self-efficacy was a robust and significant predictor of vaccination intentions at both timepoints:

• Before vaccine rollout (Time 1): OR = 1.31, 95% CI [1.02, 1.70], p < .05
• After vaccine rollout (Time 2): OR = 1.36, 95% CI [1.06, 1.75], p < .05

These results indicate that individuals with greater confidence in their ability to protect themselves against COVID-19 also showed stronger vaccination intentions.

Perceived Severity: Perceived severity was consistently significant as a predictor of vaccination intentions at both timepoints:

• Before rollout: OR = 1.18, 95% CI [1.11, 1.25], p < .001
• After rollout: OR = 1.21, 95% CI [1.14, 1.28], p < .001

Individuals who perceived COVID-19 as more severe had greater vaccination intentions, a finding aligned with the Health Belief Model.

Perceived Susceptibility: This predictor showed a differential pattern:

• Before vaccine rollout: OR = 0.91, 95% CI [0.86, 0.97], p < .001 (significant negative association)
• After vaccine rollout: OR = 0.96, 95% CI [0.90, 1.02], p = .164 (not significant)

The negative association before rollout was unexpected, suggesting that greater perceived susceptibility was associated with weaker vaccination intentions before rollout. However, this association disappeared after rollout.

Psychological Variables: Importantly, psychological variables (perceived stress, psychological flexibility, positive affect, and prosociality) were NOT significant predictors of vaccination intentions. This finding contrasts with what might intuitively be expected and suggests these variables may influence vaccination indirectly, perhaps through specific cognitive appraisals (such as perceived susceptibility and severity).

Sociodemographic Differences: There were no significant differences in vaccination intentions across sociodemographic categories.

Discussion and Conclusions

This study makes several important contributions to understanding predictors of COVID-19 vaccination intentions:

Robustness of Socio-Cognitive Predictors: The most prominent findings suggest that self-efficacy and perceived severity are robust and consistent predictors of vaccination intentions, both before and after vaccine rollout. This highlights the importance of cognitive predictors in health decisions, consistent with theoretical frameworks such as the Health Belief Model and risk behavior theories.

The Paradoxical Role of Perceived Susceptibility: The finding that perceived susceptibility was negatively associated with vaccination intentions before rollout is intriguing and counterintuitive. The authors suggest this might reflect complex dynamics: during the pre-rollout period, when vaccines were unavailable, people perceiving themselves as more susceptible might have experienced a sense of helplessness or resignation, reducing their intention to be vaccinated (a procedure not yet available). After rollout, when vaccines were available, this association disappeared, suggesting that the availability of a protective strategy (the vaccine) changed the relationship between perceived susceptibility and vaccination intention.

Psychological Variables as Indirect Predictors: The fact that psychological variables such as stress, psychological flexibility, positive affect, and prosociality were not significant direct predictors of vaccination intentions is important. The authors suggest these variables might operate indirectly, shaping specific cognitive appraisals (such as perceptions of severity and susceptibility) that in turn predict vaccination intentions. This finding points to the need for more complex models considering indirect pathways of influence.

Cultural and Policy Contexts: The multinational study with data from 33 countries suggests these prediction patterns remain relatively consistent across diverse cultural contexts. However, the authors acknowledge that country-level controls were not explicitly examined, which could have moderated these associations.

Practical Implications: The findings suggest interventions designed to increase vaccination intentions could benefit from focusing on strengthening self-efficacy for protection and providing clear information about COVID-19 severity, rather than emphasizing individual risk or susceptibility alone.

Recognized Limitations: The authors identify several important limitations: (1) the absence of country-level controls, which could have revealed important variations according to political and public health contexts; (2) the predominantly female composition of the sample (83.1% women), which limits generalizability; (3) the self-reported nature of the data; and (4) the fact that vaccination intention was measured only at Time 2, preventing examination of changes in intentions over time.

Importancia y Contribución

Este estudio contribuye significativamente al campo de la investigación en vacunación y comportamiento en salud al identificar que los predictores socio-cognitivos (autoeficacia y severidad percibida) son más robustos que los factores psicológicos emocionales generales en la predicción de intenciones de vacunación contra COVID-19. Especialmente, el análisis longitudinal multinacional demuestra cómo el contexto disponible para la acción (disponibilidad de vacunas) modula la relación entre evaluaciones de riesgo percibido e intención conductual, sugiriendo que los modelos del comportamiento de salud deben considerar explícitamente la disponibilidad de estrategias de protección. El alcance multinacional proporciona evidencia de que estos patrones se generalizan más allá de contextos culturales específicos.

---

Significance and Contribution

This study contributes significantly to the field of vaccination and health behavior research in several ways. First, it is the first multinational study explicitly examining both emotional and socio-cognitive predictors of COVID-19 vaccination intention during two critical moments differentiated by vaccine rollout. This temporal perspective is crucial for understanding how factors influencing vaccination decisions change dynamically in relation to significant contextual shifts.

Second, the study provides solid evidence that socio-cognitive predictors (self-efficacy and perceived severity) are more robust than emotional psychological factors in predicting vaccination intentions. This finding has important theoretical implications, suggesting that models emphasizing specific cognitive appraisals may be more effective than those centered on general emotional states for understanding vaccination decisions.

Third, the finding that psychological variables such as psychological flexibility, prosociality, stress, and positive affect are not significant direct predictors opens new research directions. It suggests the relationship between general psychological well-being and vaccination behavior may be more complex than previously assumed, potentially operating through specific cognitive mechanisms.

Fourth, the differential pattern observed for perceived susceptibility (negative before rollout, non-significant after) illustrates how available context for action (vaccine access) modulates the relationship between perceived risk evaluations and behavioral intention. This has implications for health behavior theory, suggesting models must explicitly consider the availability of protective actions as moderators of these relationships.

Finally, the multinational scope of the study, including data from 33 countries, provides evidence these prediction patterns generalize beyond specific cultural contexts. Although the study did not explicitly control for country-level variables, the consistency of findings suggests the cognitive mechanisms driving vaccination intentions may be universal, though potentially modulated by cultural and political contexts.

Significance and Contribution

This study contributes significantly to the field of vaccination research and health behavior by identifying that socio-cognitive predictors (self-efficacy and perceived severity) are more robust than general emotional psychological factors in predicting COVID-19 vaccination intentions. Particularly, the longitudinal multinational analysis demonstrates how the context available for action (vaccine availability) modulates the relationship between perceived risk assessments and behavioral intention, suggesting that health behavior models should explicitly consider the availability of protective strategies. The multinational scope provides evidence that these patterns generalize beyond specific cultural contexts.

---