Phenomenology of Positivity: Understanding the Joint Dynamics of Self-esteem, Life Satisfaction, and Optimism

The unique aspect of Caprara and colleagues’ conceptualization of positivity is the emphasis on the basic attitude, or the stance that self-reflective agents adopt towards their existence. This perspective differs from other models grounded in alternative theoretical premises concerning human predispositions for addressing the fundamental tasks of existence, personality organization, and the development of self-evaluative dispositions. Diener et al. (2000) pointed to a positive disposition and developed specific measures to assess life satisfaction, flourishing, positive and negative feelings, and positive thinking. However, life satisfaction has consistently been kept apart from self-esteem and optimism, which have been traced to a common “flourishing” dimension, together with purpose and respect. Huppert and So’s (2013) multidimensional model of flourishing incorporates ten dimensions: competence, emotional stability, engagement, meaning, optimism, positive emotion, positive relationships, resilience, self-esteem, and vitality.

In contrast, the positivity model uniquely traces self-esteem, optimism, and life satisfaction to a common latent factor, with life satisfaction positioned as a determinant rather than a mere indicator of happiness. Fredrickson’s (2001) broaden-and-build theory identifies positive emotions as a key resource for personal growth, but viewing positivity as a basic predisposition is different from identifying positivity with positive feelings. Although positive affect and emotions are strongly associated with positivity, the former cannot subsume the latter. Positivity, indeed, is conceptualized as a basic predisposition rather than as equivalent to the experience of positive emotions. Positive affect and emotions contribute to sustaining positivity. Yet, it is the view that individuals hold about themselves and life as worthwhile, as well as their expectations regarding desirable outcomes, that largely shape the conditions for experiencing and benefiting from positive emotions (Alessandri et al., 2014; Caprara et al., 2017a, 2024).

Positing positivity as a disposition towards existence that exerts a pervasive impact on thought, motivation, and action is consistent with the view of humans as agents with the potential to reflect upon their experiences, appropriate their thoughts, feelings, and actions, and make sense of their lives. This aligns with a view of personality that posits at its core an internal psychological organization, namely the self-system, which, over the course of development, accounts for the experience of being “I” and “me”, for the sense of ownership individuals hold over their thoughts, feelings, purposes, and actions, and for the sense of unity, coherence, and continuity that is unique to human existence (Caprara, 2024; Bandura, 2006; Blasi, 1991; McAdams, 2013; McAdams et al., 2021).

Its roots may eventually lie in a basic predisposition to approach and value events and experiences in the absence of adverse consequences, consistent with Cacioppo and colleagues’ concept of the positivity offset. This predisposition to assign a positive valence to neutral stimuli likely facilitates humans’ readiness to embrace novel experiences, even in the absence of external incentives. Over the course of development, this predisposition may evolve into a disposition that enables individuals to navigate their limitations and persist in the face of the challenges and adversities inherent in human existence.

Ultimately, positivity would enable humans to make sense of their existence and to preserve their attachment to life, despite its inherent decay. To this end, the concertation of feelings and beliefs that substantiate the experience of being worthy of value, being satisfied with one’s own life, and facing the future with hope is essential, although not something that is present at birth, but rather an emergent property of human beings. Findings have shown that positivity exhibits characteristics of stability and generalizability that are typically associated with basic traits (Alessandri et al., 2012; Caprara et al., 2017a, 2024). However, it differs from other basic dispositions, such as those commonly identified under the Five-Factor Model, which includes the so-called Big Five: extraversion, agreeableness, conscientiousness, emotional stability and openness to experience (Digman, 1990). In contrast to the Big Five, which concern dispositions that enable individuals to navigate fundamental tasks associated with their interactions with the physical environment and with other human beings (Digman, 1997), positivity represents a basic attitude that is necessary to cope with major challenges of the human condition, such as illness, aging, and death. Previous findings have demonstrated both the distinctions between the Big Five traits and Positivity, as well as the concerted action they may exert in the service of adaptation and well-being (Zuffianò et al., 2023). Whereas the Big Five personality traits are enduring behavioral tendencies largely rooted in early predispositions, positivity represents an enduring evaluative tendency concerning how individuals appraise, view, and construe their experiences, emerging later in development. As positivity emerges and gradually establishes itself as a pervasive attitude that individuals hold towards themselves and their lives, it significantly impacts their overall experience, inclinations, and habitual behaviors. This is due to the gradual construction of psychological structures and systems that exhibit a high degree of flexibility in how individuals manage their lives. It is likely that positivity shares much in common with subjective experiences previously addressed under constructs such as basic trust (Erikson, 1963) and secure attachment (Bowlby, 1988). In this regard, it is not possible to argue that early experiences of care and acceptance may exert long-lasting and beneficial effects on the views people hold of themselves and their life courses. Nevertheless, it remains difficult to assess the impact of early experiences on positivity, or precursors and early forms of positivity on personality functioning, until cognitive development enables children to recognize themselves, discern and report their feelings, and construct scenarios about the past and the future. Recent findings have pointed to the association of positivity with perceived parental warmth and satisfaction of relatedness needs during the transition from late childhood to early adolescence (Wang et al., 2023). Longitudinal studies of young adults have documented strong associations between positive orientation and interpersonal styles assessed according to a current version of attachment theory, supporting the model that assigns a significant role to positivity in counteracting dysfunctional anxious and avoidant relational styles (Castellani et al., 2016). It is likely that positivity is functional in maintaining one’s attachment to life in the face of challenges and despite the unavoidable setbacks and losses it may bring, although differently, over the course of life.

Most findings support the impact of positivity across various domains of functioning, from adolescence to old age, by promoting rewarding interpersonal relationships at school and at work, enhancing work performance and achievements, protecting health, and moderating the undesirable consequences of severe illness and stressful events (Caprara et al., 2016, 2017a, b; Consiglio et al., 2021, Livi et al., 2015; Luengo Kanacri et al., 2017; Sette et al., 2022; Tabernero Urbieta et al., 2021); Thartori et al., 2021; Theodoru et al., 2021, Zuffianò et al., 2023). Moreover, many studies have demonstrated the validity of positivity across cultures and domains. For example, Tian and colleagues (2018) noted that the Positivity scale has been used in research in America, Canada, Japan, Italy, Germany, Spain, Poland, Serbia, and Brazil, and validated themselves the scale among Chinese adults and adolescents, reporting good construct validity, high internal consistency, and measurement invariance across gender. Also the French version showed satisfactory internal consistency and test-retest stability (Vancappel et al., 2022), and the authors suggest that Positivity appears to be a common psychological resource across cultures. Turkish validation confirmed the scale structure, reported satisfactory reliability and measurement invariance, and found expected correlations with self-esteem, optimism, pessimism, life satisfaction, and affect (Duy & Yıldız, 2019).

It is likely that Positivity functions in ways similar to those attributed to self-enhancement bias and motives, self-serving cognitive distortions, and positive illusions (Alicke & Sedikides, 2011; Cummins, and Nistico, 2002; Krueger, 1998; Taylor & Brown, 1988; Zell et al., 2020; Caprara et al., 2013). Indeed, all these mechanisms are functional for enabling humans to meet the challenges of growth and to live with themselves despite their limitations. Yet, the extent to which their common functions can be traced to common or distinct motivational structures remains to be investigated.

A topic warranting further investigation, which this paper addresses, is the extent to which positivity can be considered a synergistic factor resulting from the joint effects of self-esteem, life satisfaction, and optimism, thus attesting not only to a basic, stable disposition but also to a dynamic, state-like, emerging dimension that responds adaptively to specific life challenges. While some findings indicate that the influence of positivity on desirable outcomes extends beyond that of self-esteem, life satisfaction, and optimism (Caprara et al., 2016), other findings caution against underestimating the distinct impact that self-esteem, life satisfaction, and optimism may still exert, even when positivity is taken into account (Caprara et al., 2017a, b). Equally, the reciprocal influence that self-esteem, life satisfaction, and optimism may exert on each other over the course of life and/or in the presence of particular events should not be overlooked.

As anticipated above, the present study has been conceived with the overall objective to understand to which extent the variability in positivity overlaps with variabilities in life satisfaction, optimism, and self-esteem.

Specifically, the first aim is to investigate whether positivity can be extracted from these three facets by distinguishing a between-person, trait-like component related to stable, individual differences across time, and within-person, state-like components that contribute to the explanation of the similarities in the dynamic fluctuations of self-esteem, life satisfaction, and optimism over time.

The second aim is to test whether, once positivity is accounted for, these three variables still retain their own explanatory specificity over time, thus showing that state-like positivity, as a latent within-person common core, helps to explain their temporal dynamics without subsuming them. Although previous studies (Caprara et al., 2016; Caprara, Di Giunta Caprara et al., 2017a, b) have indicated that positivity accounts for a greater proportion of the variance in outcomes than self-esteem, life satisfaction, and optimism, acknowledging the importance of positivity did not necessary imply the devaluation of the distinctive explanatory power of these other dimensions on psychological well-being.

The third aim is to evaluate whether the direct assessment of positivity through the dedicated scale effectively captures both a stable, trait-like between-person component and dynamic, state-like within-person fluctuations of the common core of self-esteem, life satisfaction, and optimism.

Ultimately, subjective mental health perception was included as a criterion to determine the extent to which positivity encompasses the explanatory contributions of self-esteem, life satisfaction, and optimism. This was achieved by observing the associations between perceived mental well-being, the within-person residuals of these constructs, and both trait-like and state-like positivity. To address these aims, the circumstances of the COVID-19 global pandemic provided a unique opportunity to examine the degree to which positivity captures the commonalities among self-esteem, life satisfaction, and optimism, and to investigate the reciprocal influences exerted by these factors both before and after an event that profoundly affected society and individuals’ lives.

The present research relies on a dataset collected within a broader teaching-innovation project¹ that explored the impact of various educational and psychological variables on students’ academic performance and study persistence in higher education. Specifically, it analyzed the relationships between variables assessed in the field of psychoeducational research, such as self-efficacy and motivation, and constructs investigated in health psychology.

The data considered in the present study were collected through a longitudinal study conducted in Spain during the global SARS-CoV-2 pandemic across three equidistant waves, each spaced approximately four months apart. The first wave occurred in January 2020, prior to the global onset of the COVID-19 pandemic. The second wave was administered in June 2020, during the peak of the pandemic when lockdowns were widely implemented across Europe to contain the spread of the virus. The final measurement wave took place in September 2020, as COVID-19 cases were declining following the summer, in response to measures implemented by various nations, and coincided with the release of initial results from COVID-19 vaccine trials.

Participants in the present study consisted of 1375 students enrolled at a mixed remote/in-person university in Spain, with an average age of 35.56 years (SD = 11.11) at the first time point of measurement². Of all participants, 26.84% were male. In January 2020, students were initially invited to participate in a longitudinal study, which included a variety of questionnaires aimed at exploring topics of relevant interest for advancing psychological knowledge.

After providing informed consent, participants completed the surveys using a secure online assessment tool via the Qualtrics website (http://www.qualtrics.com). The study protocols were approved by the UNED Ethics Committee.

Of all the participants who responded to the questionnaire at the first time point (hence, T1), the participation rate decreased to 86.18% (N = 1185) at the second time point (T2), and to 75.85% (N = 1043) at the third time point (T3). To assess potential attrition bias, baseline characteristics between participants who remained in the study and those who dropped out were analyzed. First, a chi-square test revealed no significant association between gender and attrition status between T1 and T3: χ²(1) = 0.39, p = 0.531. An independent samples t-test was conducted to compare the baseline age of participants who completed the questionnaire at the last measurement occasion and those who did not, showing that completers were slightly older (M = 36.09) compared to dropouts (M = 33.88): t(571.36) = 3.21, p = 0.001. Furthermore, subsequent independent samples t-tests showed no significant differences between completers and dropouts at T3 in baseline self-esteem (t(555.43) = 1.78, p = 0.075), optimism (t(552.31) = 0.77, p = 0.439), life satisfaction (t(562.68) = 0.23, p = 0.819), and positivity (t(548.91) = 0.66, p = 0.508).

Descriptive statistics and zero-order correlations are presented in Appendix Table; the variables show skewness and kurtosis values within acceptable limits, specifically between − 2 and + 2 (Trochim & Donnelly, 2006).

Nevertheless, Mardia’s test for multivariate normality (1970) showed significant departures from normality for life satisfaction, optimism, and self-esteem at T1 (skewness: 223.18, p < 0.001; kurtosis: 8.21, p < 0.001), T2 (skewness: 239.42, p < 0.001; kurtosis: 18.58, p < 0.001), and T3 (skewness: 190.14, p < 0.001; kurtosis: 7.60, p < 0.001). Similarly, Shapiro-Wilk’s test for univariate normality (1965) rejected the null hypothesis of normality for the Positivity scores at each occasion (T1: W = 0.96, p < 0.001; T2: W = 0.96, p < 0.001; T3: W = 0.96, p < 0.001). Consequently, the MLR robust estimator in Mplus was used to correct for non-normality in the estimation procedures.

In addition, Little’s test (1988) indicated that the missing data were produced by an MCAR mechanism: χ²(164) = 163.00, p = 0.504, allowing the use of Full Information Maximum Likelihood (FIML) estimation for handling of missing data, which under the (less restrictive) MAR assumption uses all available data points to estimate model parameters without introducing significant biases in parameter estimates and standard errors (Enders & Bandalos, 2001).

To facilitate comparisons across the constructs, all scale scores were also given in Appendix Table as Percentages of Scale Maximum (%SM), ranging from 0 to 100%. The mean values for all constructs across the three time points were consistently above the neutral threshold of 50% SM (Cummins & Nistico, 2002) (Fig. 1).

To model how the constructs evolved across three occasions (before, during, and after the COVID-19 pandemic), four separate univariate multilevel (mixed-effects) models were fitted, estimated by full-information maximum likelihood to retain all available cases under the missing-at-random assumption (Enders & Bandalos, 2001). The fixed effects component included linear and quadratic effects of time, while the random part allowed intercepts and slopes to vary by subject. The results are presented in Table 1.

Fixed-effect estimates indicated that, relative to the pre-pandemic assessment, self-esteem, optimism, and positivity declined during COVID-19 and rebounded after the pandemic. Life satisfaction increased modestly across occasions. Although statistically significant, time effects accounted for no more than 0.5% of the total variance in any outcome.

After the comparison of the first two models to evaluate the invariance of within-person factor loadings of positivity across time, the chi-square difference test showed that the constrained model demonstrated significantly poorer fit relative to the unconstrained model (Δχ²(4) = 17.48, p = 0.001), thus supporting the preference for the less-restrictive model. Consequently, these factor loadings were allowed to vary over time in all subsequent analyses.

On the contrary, when the time-invariance of the within-person effects was tested, the chi-square difference test did not show a significant improvement in model fit for the unconstrained model compared to the constrained model: Δχ²(9) = 13.87, p = 0.127. Therefore, the more parsimonious constrained model was retained as the best-fitting model: χ² = 20.07, df = 12, p = 0.066; CFI = 0.99; TLI = 0.99; RMSEA = 0.02, 90% CI [0.000, 0.039], p = 0.999; SRMR = 0.03. When the selected model was tested with gender and age as covariates, the structural relationships remained unchanged (Fig. 2).

From the results it can be seen how the latent trait-like positivity factor showed strong standardized loadings onto the random intercepts of self-esteem (λ = 0.85, p ≤ 0.001), life satisfaction (λ = 0.88, p ≤ 0.001), and optimism (λ = 0.95, p ≤ 0.001), indicating that trait-like positivity accounts for a relevant proportion of the between-person variance in each construct. The particularly high loading for optimism suggests that most of its between-person variance is shared with the underlying trait-like positivity factor. Regarding the within-person fluctuations, the latent state-like positivity factors at each time point demonstrated significant standardized loadings onto the within-person deviations of the three constructs, ranging from λ = 0.30, p ≤ 0.001 (between state-like positivity and life satisfaction at T2), to λ = 0.69, p ≤ 0.001 (between state-like positivity and optimism at T2). These findings indicate that state-like positivity captures a significant portion of the shared within-person variability in self-esteem, life satisfaction, and optimism over time.

Regarding the within-person residuals of self-esteem, life satisfaction, and optimism, that is, the time-specific deviations unexplained by the latent state-like positivity factor, the results demonstrate significant temporal dynamics (Fig. 3). Particularly, each residual showed significant carry-over effects, indicating that fluctuations in a variable’s residual at one time point predicted fluctuations in the same residual at the subsequent time point. Significant spill-over effects were observed among the residuals of all the variables, with the exception of the spill-over effect from self-esteem to life satisfaction, which was not significant. Overall, these results demonstrate that the within-person residuals maintain meaningful temporal relationships independent of the state-like positivity factor. In other words, the state-like positivity factor does not fully overlap with the informative variability of the three constructs at the within-person level, indicating that there are substantive aspects of self-esteem, life satisfaction, and optimism fluctuations that are not captured by the common state-like positivity component.

Regarding the significance of the covariates, gender marginally influenced only self-esteem at T1 (β = −0.05, p = 0.046). In contrast, age demonstrated a significant impact across all variables and occasions, affecting life satisfaction measured at T1 (β = 0.16, p < 0.001), T2 (β = 0.14, p < 0.001), and T3 (β = 0.12, p < 0.001), optimism at T1 (β = 0.23, p < 0.001), T2 (β = 0.19, p < 0.001), and T3 (β = 0.20, p < 0.001), and self-esteem at T1 (β = 0.28, p < 0.001), T2 (β = 0.28, p < 0.001), and T3 (β = 0.27, p < 0.001).

The variance of the observed P-scores was decomposed into a component related to stable time-invariant individual differences in measured Positivity, and a component related to its within-person fluctuations. Table 2 shows the zero-order correlations between the trait-like between-person component obtained from the RI-CLPM, considering only self-esteem, life satisfaction, and optimism, on one side, and the same between-person component derived from the multilevel decomposition of the directly measured Positivity scores. Similarly, the correlations between the factor scores of the within-person temporal fluctuations in Positivity from the RI-CLPM, and the corresponding within-person components extracted from the observed Positivity scores at each measurement occasion are shown. These correlations assess the convergence, respectively, of the trait-like and state-like components obtained through the two methodologies applied to the different scales, indicating the extent to which they capture the same underlying constructs.

First, the high correlation (r = 0.92, p < 0.001) between the trait-like between-person component of positivity extracted from the RI-CLPM and the measured between-person component of positivity using the P-scale shows that the P-scale effectively captures the more stable, trait-like aspect of positivity. This indicates that the P-scale’s between-person variance substantially represents the common variability in the stable individual differences of self-esteem, life satisfaction, and optimism. Moreover, the significant positive correlations between the within-person fluctuations of the P-scores and the corresponding within-person positivity fluctuations from the RI-CLPM at each time point (T1, T2, and T3) indicate that the P-scale also captures the state-like components of positivity, which reflects the contemporaneous fluctuations in self-esteem, life satisfaction, and optimism. This alignment shows that the P-scale is also sensitive to the within-person changes that occur jointly in self-esteem, life satisfaction, and optimism, thus capturing the dynamic aspects of positivity alongside its stable traits.

Lastly, Table 3 shows that subjective mental health perception is strongly related to overall positivity (P-score), while also correlating substantially with the between-person and within-person positivity components obtained from the RI-CLPM: these results indicate that both stable, trait-like differences and more transient, state-like fluctuations in positivity (simultaneously captured by the P-score) bear meaningful associations with individuals’ subjective mental health. Moreover, correlations of comparable magnitude emerge between subjective mental health perception and the residual within-person fluctuations in self-esteem, life satisfaction, and optimism, even after accounting for the common state-like positivity factor. Therefore, despite the shared influence of state-like positivity, these three constructs retain independent explanatory value and are not fully subsumed by either trait- or state-like positivity, as time-sensitive deviations in self-esteem, life satisfaction, and optimism continue to play a significant role in relation to mental well-being.

The reproducible code plan (R Markdown), datasets, knitted files, Mplus inputs and outputs are available on the Open Science Framework (OSF) at the following link: https://osf.io/xwu6y.