Original quantitative research – Reciprocal associations between positive emotions and resilience predict flourishing among adolescents

Table 2. Descriptive statistics and bivariate correlations with corresponding confidence intervals between study variables

Variable	Mean (SD)	1. Positive Emotions – Wave 1	2. Positive Emotions – Wave 2	3. Positive Emotions – Wave 3	4. Resilience – Wave 1	5. Resilience – Wave 2	6. Resilience – Wave 3
1. Positive Emotions – Wave 1	2.76 (0.88)	Same category	N/A	N/A	N/A	N/A	N/A
2. Positive Emotions – Wave 2	2.74 (0.87)	0.42 (0.39–0.44)	Same category	N/A	N/A	N/A	N/A
3. Positive Emotions – Wave 3	2.69 (0.88)	0.35 (0.31–0.38)	0.42 (0.39–0.45)	Same category	N/A	N/A	N/A
4. Resilience – Wave 1	3.64 (0.98)	0.26 (0.23–0.29)	0.23 (0.19–0.25)	0.20 (0.16–0.23)	Same category	N/A	N/A
5. Resilience – Wave 2	3.58 (1.00)	0.22 (0.18–0.25)	0.30 (0.27–0.32)	0.24 (0.21–0.28)	0.46 (0.43–0.48)	Same category	N/A
6. Resilience – Wave 3	3.54 (1.02)	0.17 (0.14–0.21)	0.22 (0.19–0.26)	0.33 (0.29–0.36)	0.39 (0.35–0.42)	0.49 (0.46–0.52)	Same category
7. Flourishing – Wave 3	33.16 (12.64)	0.09 (0.05–0.13)	0.14 (0.10–0.18)	0.27 (0.24–0.31)	0.18 (0.14–0.21)	0.17 (0.13–0.21)	0.35 (0.32–0.38)
Abbreviation: SD, standard deviation.
Notes: Wave 1 = 2017/18 school year; Wave 2 = 2018/19 school year; Wave 3 = October–December 2019, as only those participants who completed the questionnaire in the first half of the school year were included. Confidence intervals that do not include zero are indicative of significant correlations.

The intraclass correlations for positive emotions and resilience across the three waves were .47 and .52, respectively, indicating that approximately half of the variability in positive emotions and resilience is attributable to stable between-person differences and half is attributable to within-person variation. When the clustering of students within schools was taken into account, the intraclass correlations for positive emotions, resilience and flourishing were.04,.03 and.01, respectively, indicating that this clustering contributed little variability in the scores. As such, the clustering of schools was not considered in the analyses.

Longitudinal measurement invariance of positive emotions was supported. A chi-square test comparing a constrained model in which the factor loadings were identical over time and an unconstrained model in which the factor loadings were free to vary was not significant (p > 0.05), suggesting that the latent structure is invariant over time.

Both a CLPM and RI-CLPM were fit to the data to determine that a RI-CLPM was indeed the most parsimonious model. Inspection of the fit indices provided support for the RI-CLPM (aBIC = 77 090; CFI =.997; RMSEA =.02) being the better fit to the data than the CLPM (aBIC = 77 287; CFI =.948; RMSEA =.08). A model with the autoregressive and cross-lag parameters freely estimated failed to converge. As a result, both the autoregressive and cross-lagged parameters were constrained to be invariant over time. The specified model showed excellent fit to the data (χ2(7) = 18.29, p = 0.011, aBIC = 77 090.43; RMSEA =.02; CFI =.99; SRMR =.01).

Standardized estimates of the RI-CLPM are shown in Figure 1.

Figure 1 - Text description

Figure 1 illustrates standardized results of RI-CLPM between positive emotions and resilience predicting flourishing. Measurement error refers to the variability that remains unaccounted for in an indicator's measurement of its underlying construct. In the diagram, this is represented by a circle with a one-way arrow.

Association between variables	β
Between-person level association
Positive emotion (Between) to Resilience (Between)	0.44
Resilience (Between) to Positive emotion (Between)	0.44
Positive emotion (Between) to Flourishing Wave 3	0.08
Resilience (Between) to Flourishing Wave 3	0.19
Within-person autoregressive association
Positive emotion Wave 1 (Within) to Positive emotion Wave 2 (Within)	0.11
Positive emotion Wave 2 (Within) to Positive emotion Wave 3 (Within)	0.11
Resilience Wave 1 (Within) to Resilience Wave 2 (Within)	0.17
Resilience Wave 2 (Within) to Resilience Wave 3 (Within)	0.17
Cross-lagged association
Resilience Wave 1 (Within) to Positive emotion Wave 2 (Within)	0.07
Resilience Wave 2 (Within) to Positive emotion Wave 3 (Within)	0.07
Positive emotion Wave 1 (Within) to Resilience Wave 2 (Within)	0.10
Positive emotion Wave 2 (Within) to Resilience Wave 3 (Within)	0.10
Within-person contemporaneous association
Positive emotion Wave 1 (Within) to Resilience Wave 1 (Within)	0.16
Resilience Wave 1 (Within) to Positive emotion Wave 1 (Within)	0.16
Positive emotion Wave 2 (Within) to Resilience Wave 2 (Within) (with measurement error)	0.20
Resilience Wave 2 (Within) to Positive emotion Wave 2 (Within) (with measurement error)	0.20
Positive emotion Wave 3 (Within) to Resilience Wave 3 (Within) (with measurement error)	0.26
Resilience Wave 3 (Within) to Positive emotion Wave 3 (Within) (with measurement error)	0.26
Temporal deviations
Positive emotions Wave 3 (Within) to Flourishing Wave 3	0.17
Resilience Wave 3 (Within) to Flourishing Wave 3	0.25

At the between-person level, positive emotions and resilience were positively associated (β=.44; 95% CI = 0.36–0.53). In other words, on average, people who reported more positive emotions also reported greater resilience. There was a significant within-person contemporaneous association between positive emotions and resilience in 2017/18 (β = .16; 95% CI = 0.10–0.21), 2018/19 (β =.20; 95% CI = 0.15–0.26) and 2019 (β =.26; 95% CI = 0.21–0.31). There were significant within-person autoregressive associations for both positive emotions (β = .11; 95% CI = 0.05–0.17) and resilience (β =.17; 95% CI = 0.10–0.22).

Of primary interest are the cross-lagged relationships. We examined the cross-lagged relationship between adolescents' positive emotions at a given time point and their resilience at the subsequent time point to test the premise that positive emotions function to build resources over time and that greater accrual of personal resources should lead to more positive emotions. The findings supported these suppositions. Greater positive emotions positively predicted resilience at the subsequent wave (β =.07; 95% CI = 0.02–0.12). Cross-lagged associations were also supported from resilience and positive emotions at the subsequent wave (β =.10; 95% CI = 0.05–0.15).

Finally, temporal deviations in both positive emotions (β =.17; 95% CI = 0.11–0.23) and resilience (β =.25; 95% CI = 0.20–0.31) at Wave 3 predicted variation in flourishing at Wave 3. Stable between-person differences in positive emotions (β =.08; 95% CI = 0.02–0.14) and resilience (β =.19; 95% CI = 0.13–0.25) also significantly predicted variation in flourishing. Overall, the model accounted for 17% of the variance in flourishing.

Discussion

Since the broaden and build theory was introduced in 1998, nearly all theoretical tests have examined between-person differences.^{Footnote 3} As a result, the current state of knowledge is limited to an understanding of individual differences in this process. Our aim in this study was to examine the upward spiral towards well-being by examining both between- and within-person associations.

Our findings support our hypothesis. When individuals experienced more positive emotions than they usually do, they experienced an increase in resilience one year later. Similarly, higher levels of resilience were subsequently associated with more frequent experiences of positive emotions. Higher levels of resilience and of positive emotions predicted enhanced flourishing. At the between-person level, both positive emotions and resilience were positively related and predicted greater flourishing.

Our results are consistent with the central premise of the broaden and build theory that experiences of positive emotions, even if fleeting, build enduring psychological resources and trigger upward spirals towards enhanced well-being. By examining within-person associations, the results of the current study provide initial support for the causal process hypothesized in the broaden and build theory of positive emotions. Our results also provide a deeper and more robust understanding of the upward spiral and important theoretical and applied contributions to the study of adolescent well-being.

The upward spiral dynamic towards increased well-being is best captured by prospective, reciprocal associations. Although researchers have previously examined experimental and prospective associations between positive emotions and resilience, this is the first study to examine the upward spiral between positive emotions and resilience. Unlike prospective observational studies, tests of reciprocal relationships can provide stronger evidence of the upward spiral effect by demonstrating that positive emotions and resilience mutually reinforce one another to enhance well-being, and that this is a process that unfolds within individuals over time.

At the between-person level, positive emotions and resilience were moderately positively associated. This finding is consistent with prior studies that also found relationships between positive emotions and resilience of similar magnitudes when assessed at the between-person level.^{Footnote 9 Footnote 10 Footnote 11 Footnote 28} In the present study, we found smaller associations between positive emotions and resilience when these were assessed both contemporaneously and prospectively at the within-person level. The difference in the magnitude of these relationships is in line with Burns et al.,^{Footnote 27} who found that the CLPM results are overestimated relative to RI-CLPM results. Smaller estimates are expected relative to prior estimates as autoregressive and lagged relationships pertain only to within-person associations in the RI-CLPM, while the CLPM does not separate out between-person differences from the within-person associations.^{Footnote 22}

Although effect sizes are smaller than those reported in previous studies employing between-person designs, dismissing small effect sizes has been cautioned against.^{Footnote 29} Small effect sizes can yield important theoretical advances and hold considerable impact. This is particularly true for effects that accumulate or compound over time, such as the reciprocal nature of the associations we documented in this study.^{Footnote 30}

This study supports not only the broaden and build theory but also the upward spiral dynamic;^{Footnote 4 Footnote 5} positive emotions were associated with increased resilience one year later, illustrating that positive emotions function to build personal resources (i.e. resilience). Furthermore, experiences of positive emotions and resilience were mutually reinforcing and predicted increased flourishing across three years, illustrating the upward spiral dynamic. These findings are notable given that participants reported their positive emotions over the previous two weeks and resilience was assessed one year later. On average, participants reported experiencing positive emotions a few times in the previous two weeks. That the meaningful variance was accounted for is notable, particularly in the absence of any intervention and the partitioning of variance into within- and between-person variance, which leaves less variance to predict year-to-year.

Schools are one of the most important developmental contexts in adolescents' lives. Schools afford key opportunities for adolescents to develop the skills and competencies that support their capacity for successful development into adulthood. Coinciding with the rise of positive psychology, many schools have begun shifting from focussing on negative experiences and deficits to prioritizing the implications of positive school experiences on student flourishing.^{Footnote 31 Footnote 32} Positive psychology interventions implemented in schools have demonstrated promise for promoting both positive emotions and resilience among youth.^{Footnote 33 Footnote 34} The findings of this study could be integrated into existing positive education programs in schools to promote optimal functioning among youth.

Strengths and limitations

Although we assessed only a limited number of positive emotions, feelings of hope are thought to be the positive emotion most central to resilience processes. Hope is theorized to build resilience, and early empirical research substantiates this claim.^{Footnote 3 Footnote 35 Footnote 36 Footnote 37}

Another limitation of this study is that resilience was assessed using a single indicator. As a result, the ability to capture the full breadth of the construct is limited. The decision to use a single-item indicator of resilience was made in response to concerns about participant burden, survey space and adequate time to complete the survey in the time provided by the schools.

Only data collected in the fall of 2019 were used in Wave 3. This may result in biased estimates given the noted seasonal differences in mental health and well-being in Canada. The statistical approach used to model the longitudinal data has limitations, including the assumption of linear relationships between time-invariant variables and the inability to capture more complex processes.

Finally, since the schools analyzed in the study were chosen for convenience, the findings cannot be generalized beyond the sample. Nevertheless, by implementing a passive-consent protocol, the study was able to achieve high participation rates and minimize selection bias within schools. Moreover, because of the large sample size, the conclusions drawn from this research are likely to be applicable to a considerable number of high school students in Canada.

Several strengths of the current study should also be considered. This is the first study to examine the upward spiral towards enhanced well-being in a sample of adolescents. The longitudinal design allowed for the investigation of reciprocal relationships and for the disaggregation in variance at the between and within-person levels as well as sufficient time for resources to accrue and affect well-being. This allowed for a more accurate test of the broaden and build theory, and responds to a recent call for researchers to employ more advanced statistical modelling when evaluating the broaden and build theory to allow for disaggregation in within- and between-person variance.^{Footnote 23}

Conclusion

Much of the research on adolescent mental health has adopted a pathological approach, investigating mental illness at the expense of understanding factors that lead to optimal psychological functioning. Estimates indicate that flourishing decreases across adolescence.^{Footnote 2} Strategies that can mitigate this decline need to be identified. The findings from this study provide evidence that even fleeting experiences of positive emotions are associated with increases in resilience one year later, which subsequently result in further increases in positive emotions. These experiences catalyze an upward spiral towards flourishing. Opportunities to enhance emotions early on in adolescence may help build resources that can set students on the path towards increased well-being.

Acknowledgements

The COMPASS study has been supported by a bridge grant from the Canadian Institutes of Health Research (CIHR) Institute of Nutrition, Metabolism and Diabetes (INMD) through the "Obesity – Interventions to Prevent or Treat" priority funding awards (OOP-110788; awarded to SL); an operating grant from the CIHR Institute of Population and Public Health (IPPH) (MOP-114875; awarded to SL); a CIHR project grant (PJT-148562; awarded to SL); a CIHR bridge grant (PJT-149092; awarded to KP/SL); a CIHR project grant (PJT-159693; awarded to KP); and a research funding arrangement with Health Canada (#1617-HQ-000012; contract awarded to SL).

The COMPASS-Quebec project also benefits from funding from the Quebec Ministère de la Santé et des Services sociaux and the Direction régionale de Santé Publique du CIUSSS de la Capitale-Nationale. This research was supported by the Social Sciences and Humanities Research Council of Canada through a postdoctoral fellowship awarded to JG. The data and materials that support the findings of this study are available from the corresponding author upon reasonable request. Comprehensive code used for analyses has been made publicly available and can be accessed at osf.io/jueg8.

Authors' contributions and statement

JG: Conceptualization, Methodology, Writing – Original draft, Writing – Review & Editing.

MG: Data curation, Formal analysis, Visualization.

LB: Conceptualization, Writing – Review & editing.

KP: Funding acquisition, Investigation, Project administration, Resources, Writing – Review & Editing.

SL: Funding acquisition, Project administration, Resources, Supervision, Writing – Review & Editing.

The content and views expressed herein are those of the authors and do not necessarily reflect those of the Government of Canada.

Conflicts of interest

None.

Table of Contents | Next