Developmental Psychology
Equitable Shifts in Youth Resilience? Distinguishing Normative Changes
and Pandemic Effects on Academic Self-Efficacy and Cognitive Reappraisal
Juuso Repo, Sanna Herkama, and Christina Salmivalli
Online First Publication, January 23, 2025. https://dx.doi.org/10.1037/dev0001913
CITATION
Repo, J., Herkama, S., & Salmivalli, C. (2025). Equitable shifts in youth resilience? Distinguishing normative
changes and pandemic effects on academic self-efficacy and cognitive reappraisal. Developmental
Psychology. Advance online publication. https://dx.doi.org/10.1037/dev0001913 
Equitable Shifts in Youth Resilience? Distinguishing Normative Changes and
Pandemic Effects on Academic Self-Efficacy and Cognitive Reappraisal
Juuso Repo, Sanna Herkama, and Christina Salmivalli
Research Flagship Center for Inequalities, Interventions and New Welfare State, University of Turku
This preregistered longitudinal study examined the long-term effects of the COVID-19 pandemic on academic
self-efficacy and cognitive reappraisal in early adolescence. It followed and compared two cohorts over 4 years:
one prepandemic (11–14 years, 2016–2019) and one during the pandemic (2019–2022). The study analyzed
annual well-being surveys merged with school enrolment data from South Australian public schools
(N = 28,307, 49% female). Employing latent growth modeling and a novel cohort comparison design, the
study addressed a major limitation in pandemic studies: It separated pandemic effects from normative
developmental changes. Results indicate that the pandemic cohort largely followed typical, yet declining,
developmental trajectories, showing resilience at a population level. Unexpectedly, the examination ofmultiple
covariates (i.e., gender, socioeconomic status, non-English background, anxiety, peer belonging, teacher
support) showed that preexisting vulnerabilities did not predict adverse pandemic effects. This research
underscores the value of longitudinal data infrastructures and the importance of understanding normative youth
development and resilience research in discerning the effects of pandemics or other widespread crises.
Public Significance Statement
This study examined how the COVID-19 pandemic affected academic self-efficacy and cognitive
reappraisal in South Australian adolescents. We compared a during-pandemic cohort (2019–2022) to a
prepandemic cohort (2016–2019) and found that most young people showed resilience during the
pandemic, maintaining typical developmental patterns. Our study underscores the importance of robust
longitudinal data and highlights the need to distinguish pandemic’s impact from normative youth
development to inform future crisis response strategies.
Keywords: adolescent, resilience, COVID-19, academic self-efficacy, cognitive reappraisal
Supplemental materials: https://doi.org/10.1037/dev0001913.supp
Gustavo Carlo served as action editor.
Juuso Repo https://orcid.org/0000-0001-8756-6936
The study design, hypotheses, and analysis plan were preregistered on the
Open Science Framework. Synthetic data replicating the structure of the original
data set, all analysis R code, software used, study materials, and synthetic data
are publicly available at https://doi.org/10.17605/OSF.IO/HKQ39.
This study’s design and analysis plan were preregistered on the Open
Science Framework. Deviations from the preregistration are reported in
the Supplemental Table 12. This article uses the Workflow for Open
Reproducible Code in Science (Van Lissa et al., 2021) to ensure
reproducibility and transparency. The preregistration, all analytic code,
and additional online material are available at https://doi.org/10.17605/
OSF.IO/HKQ39. While the data use agreement prohibits the authors from
making the full data publicly available, the authors have published
synthetic data together with the analytic code. For access to original data,
please contact the Research Unit of South Australian Department for
Education. This research was conducted in full compliance with the ethical
standards outlined in the American Psychological Association guidelines.
The procedure followed the ethical standards of the University of Turku
Ethics Committee for Human Sciences, the Finnish National Board of
Research Integrity (National Advisory Board on Research Ethics, 2009),
and the Finnish Personal Data Act (523/1999).
The authors have no conflicts of interest to disclose. The authors thank the
South Australian Department for Education for their groundbreaking work
with the Wellbeing and Engagement Data Collection and for granting
approval and access to the data for this study. The authors also thank all the
schools and students who participated in data collection. The study has been
supported financially by the Research Flagship Center for Inequalities,
Interventions and New Welfare State, University of Turku. This research
was funded in whole, or in part, by the Academy of Finland (AKA; 345546).
For the purpose of open access, the author has applied a CC BY public
copyright license to any Author Accepted Manuscript version arising from
this submission.
This work is licensed under a Creative Commons Attribution-Non
Commercial-No Derivatives 4.0 International License (CC BY-NC-ND 4.0;
https://creativecommons.org/licenses/by-nc-nd/4.0). This license permits
copying and redistributing the work in any medium or format for
noncommercial use provided the original authors and source are credited
and a link to the license is included in attribution. No derivative works are
permitted under this license.
Juuso Repo played a lead role in conceptualization, formal analysis,
methodology, visualization, and writing–original draft. Sanna Herkama
played a supporting role in supervision and writing–review and editing.
Christina Salmivalli played a lead role in supervision and a supporting role in
conceptualization and writing–review and editing.
Correspondence concerning this article should be addressed to Juuso
Repo, Research Flagship Center for Inequalities, Interventions and New
Welfare State, University of Turku, 20014 Turku, Finland. Email:
juuso.repo@utu.fi
Developmental Psychology
© 2025 The Author(s)
ISSN: 0012-1649 https://doi.org/10.1037/dev0001913
1
Regarding the pandemic’s impact on adolescents, both research
and public concern have heavily focused on (a) adverse mental
health effects and (b) growing inequalities, with harmful impacts
accumulating among those with preexisting vulnerabilities (e.g.,
Kauhanen et al., 2023; Shergold et al., 2022). Meanwhile, the
possible effects on adolescents’ self-regulation development have
been understudied, despite their centrality in psychological research,
interventions, and educational policies (Domitrovich et al., 2017;
Durlak et al., 2011; Restad & Mølstad, 2021; Smithers et al., 2018).
The pandemic’s impact on self-regulation is particularly interesting
as it may have had both positive and negative effects: detrimental for
some, yet opening new learning and development opportunities for
others (Skinner et al., 2023).
This study focuses on two key aspects of self-regulation: academic
self-efficacy (ASE) and cognitive reappraisal (CR)—a subdomain
of emotion regulation. These concepts are among the most studied in
adolescent development and are critical for academic achievement,
psychological well-being, and youth resilience (Honicke&Broadbent,
2016; McRae & Gross, 2020; Talsma et al., 2018). During the
pandemic, adolescents’ self-regulation was significantly challenged
due to the loss of in-person schooling and other disruptions to support
systems and social interactions.
Moreover, given the rise in adolescents’ internalizing symptoms
during the pandemic, and the fact that these symptoms are
associated with academic and emotional self-efficacy (Bandura,
1997; Skinner et al., 2023; Tak et al., 2017; Young et al., 2019), the
pandemic may have adversely affected the development of these
factors, disproportionately depending on individuals’ psychological
and social resources. As the increase in mental health symptoms
may reflect a normative and nonpathological response to pandemic
conditions, it is essential to assess the impact on adolescents’ self-
regulation to gain a more holistic understanding on pandemic’s
long-term effects.
Given the pandemic’s disruption of both academic and emotional
domains, this study employs a parallel analysis of these essential
areas for adolescent development. It explores the long-term impact
of the pandemic on ASE and CR in early adolescence and examines
whether preexisting vulnerabilities related to mental health, peer
relationships, and teacher relationships were exacerbated by the
crisis.
A key limitation in prior pandemic research is the failure to
separate normative developmental changes from pandemic-related
effects. Our study overcomes this gap with a longitudinal cross-
cohort design. We will begin by discussing ASE and CR within the
context of the pandemic, followed by an examination of resilience
research that highlights the complexity and individual variability in
responses to stressful life events. Additionally, we highlight some key
limitations in previous pandemic studies and outline our contribution
through four research questions.
ASE and CR
ASE, a subdomain of the broader concept of self-efficacy, represents
an individual’s judgment about their ability to successfully attain
educational goals (Bandura, 1997). Stressful academic situations, such
as struggling with schoolwork, can reduce ASE levels. By contrast,
supportive peer and teacher relationships are linked to higher ASE
(Vestad et al., 2022). Navigating the new demands of remote learning,
accompanied by shifts in students’ relational worlds with teachers
and peers, may have diminished student’s perceived competence (Fong
et al., 2022). Additionally, remote schooling may have highlighted
disparities. For example, some families were able to provide enriching
home-learning settings with strong parental support, while others,
facing socioeconomic or emotional challenges, struggled with the
transition, possibly exacerbating ASE disparities (Strasser et al., 2023).
Generally, ASE beliefs are shown to be relatively stable over
time, also in the context of the pandemic (Repo et al., in press; Talsma
et al., 2021).
CR, a subdomain of emotion regulation, is a cognitive process
involving changing one’s perspective on a situation to influence
emotional response (Gross & Thompson, 2007). Emotion regulation
is strongly associated with resilience, learning, and mental well-being
(Aldao et al., 2010; LeMoult & Gotlib, 2019; Willner et al., 2022;
Young et al., 2019). Surveys on CR usually assess self-perceived
abilities, also interpreted as self-efficacy in emotional regulation
(Alessandri et al., 2015). In adolescence, CR abilities generally
improve with cognitive and brain development (Herd et al., 2020).
However, a recent review byWillner et al. (2022) highlighted the need
for further research, as several large-scale studies have shown declines
in self-perceived CR during midadolescence (Gregory & Brinkman,
2020; Gullone et al., 2010; Vestad et al., 2022; Zimmermann &
Iwanski, 2014), particularly in girls (Sanchis-Sanchis et al., 2020).
Importantly, the pandemic has amplified emotional challenges,
especially for girls (Kauhanen et al., 2023; Madigan et al., 2023).
Recent research emphasizes CR’s interpersonal aspect, supporting
the notion that pandemic-related social disruptions combined with
emotional struggles could hinder CR development (Bandura, 1997;
Hayes et al., 2022; Hofmann, 2014; Skinner et al., 2023). Notably,
adolescents with initially lower emotion regulation have been found
to catch up in development during the school years (Herd et al.,
2020; see also Liu, 2019). However, the pandemic’s impact on this
compensation effect remains unexplored.
Ambiguities of Stress and Pandemic Effects
Current research on stress and resilience underlines significant
individual variation in stress responses, with some individuals being
more susceptible to its effects. Additionally, resilience—defined as
positive adaptation amid adversity—is not solely based on individual
characteristics but also on socioecological resources (Ungar et al.,
2013). According to the stress sensitization hypothesis (Rutter,
2012), those already vulnerable before the pandemic may have been
more affected by it, potentially intensifying preexisting disparities.
Factors linked to higher stress sensitivity in both resilience and
pandemic research include limited interpersonal support (Samji et al.,
2022), preexisting mental health issues (Alamolhoda et al., 2022),
and female gender (Madigan et al., 2023).
Conversely, many individuals, families, and communities may
have responded to the pandemic by mobilizing new capabilities,
connecting in new ways, and discovering untapped strengths to
adaptand cope with stress (Beames et al., 2021). Drawing from
the theory on stress-related growth, which pertains to the perception
or experience of deriving benefits from encountering stressful
circumstances (Vaughn et al., 2009), the pandemic may have
presented new opportunities for psychological development.
Contrary to stress sensitization, the steeling effect hypothesis
suggests that exposure to moderate stress could have supported
some adolescents’ development, cultivating resources for future
2 REPO, HERKAMA, AND SALMIVALLI
resilience (Rutter, 2012). However, these potential positive effects
may have disproportionately accumulated to those with better
preexisting socioemotional resources, indicating an exacerbating
Matthew effect.
Limitations in Previous Pandemic Research
Despite the surge in pandemic research, results have often been
ambiguous, with many studies facing notable limitations: lack of
longitudinal designs, absence of prepandemic data, emphasis on
short-term effects, and dependence on self-perceived assessments
of pandemic impact (Gorman, 2023). A less discussed but crucial
deficit is the inability to distinguish potential pandemic effects from
normative developmental changes. Very few studies have been
structured to tackle this specific challenge (Repo et al., in press;
Strohmeier & Branje, 2023; Wright et al., 2024).
Further, pandemic research has predominantly focused on
negative mental health aspects, such as anxiety, or the repercussions
of learning losses, neglecting the positive dimensions of well-being
and the potential for stress-related growth. ASE and CR have been
primarily studied as protective factors—moderating the pandemic
effects—rather than as longitudinal outcomes (see, e.g., Bagheri
Sheykhangafshe et al., 2022). Additionally, most pandemic research
exploring the connection between interpersonal resources and
psychological well-being has emphasized cross-sectional associa-
tions, even in longitudinal studies. It is clear that individuals with
positive peer relationships or more teacher support consistently
exhibit better well-being concurrently, including during the
pandemic. However, studies examining whether prepandemic levels
of these factors predicted any changes during the pandemic are
still rare.
Taken together, longitudinal studies acknowledging the complex
and varied nature of pandemic effects, addressing long-term
consequences, and differentiating normative development from
pandemic-related changes have been notably scarce or nonexistent.
This study addresses these gaps by conducting a developmental
comparison between age cohorts from prepandemic and during-
pandemic periods.
The Present Study
The current preregistered study compares the development of
ASE and CR in two same-aged early adolescent cohorts before and
during the pandemic. To our knowledge, it represents the largest
longitudinal study to date examining developmental trends in ASE
and CR across early adolescence. The primary objective is to
disentangle developmental changes from possible pandemic effects.
Moreover, the study seeks to ascertain if the pandemic exacerbated
interindividual differences in these domains and, if so, which
preexisting socioemotional factors predicted the diverse pandemic-
related responses.
To address these objectives, the study examines the following
research questions:
1. What are the typical developmental patterns observed in
ASE and CR among early adolescents?
2. To what extent did the pandemic affect the development of
ASE and CR in adolescents?
3. Did the pandemic exacerbate between-person differences
in the development of ASE and CR?
4. How much did demographic factors and initial levels of
anxiety, peer belonging, and teacher support predict
between-person differences in the development of the
outcomes? Were there any variations in these associations
between the cohorts?
The study is set in Australia, where initial success in controlling
COVID-19 was followed by a surge in infections, exceeding
double the Organisation for Economic Co-operation and
Development average (Shergold et al., 2022). Throughout the
pandemic, Australian COVID-19 policies were among the most
stringent, ranking seventh on average in the Organisation for
Economic Co-operation and Development (Hale et al., 2020). In
South Australia, where this study’s data were collected, statewide
lockdowns began in March 2020 and ended in February 2022
(Storen & Corrigan, 2020). During these periodic lockdowns,
schools were closed. Outside of the lockdowns, schools stayed open
but implemented strict social distancing and hygiene measures
(Edwards et al., 2022).
Method
Transparency and Openness
The study design, hypotheses, and analysis plan were preregis-
tered on the Open Science Framework. Synthetic data replicating the
structure of the original data set, all analysis R code, software used,
study materials, and synthetic data are publicly available at https://
doi.org/10.17605/OSF.IO/HKQ39. Access to the original data
requires approval from the South Australian Department for
Education. Any deviations from the preregistered plan are
documented in the Supplemental Materials.
Data
This study used data from two sources: an annual student well-
being survey and school enrollment data, both part of the South
Australian Wellbeing and Engagement Collection administered by
the South Australian Department for Education since 2013 (Gregory
et al., 2022). Students completed the survey using an online data
collection system during school class. Our sample includes students
from governmental schools, where about 65% of South Australian
students are enrolled (Australian Bureau of Statistics, 2022).
The sample consists of two age cohorts. The first cohort completed
Grades 6–9 before the pandemic, in 2016–2019, and the second
during the pandemic, in 2019–2022. The annual data collection took
place at different times: in 2016 (October/November), 2017 (July/
August), 2018 (July/August), 2019 (March/April), 2020 (February
and July/August), 2021 (February), and 2022 (February). The school
year in Australia runs from late January to mid-December.
The sample included 13,372 respondents in the prepandemic
cohort (49.2% female; 50.8% male; mean age at T1 = 11.2,
SD = 0.5) and 14,935 respondents in the during-pandemic cohort
(49.0% female; 51.0% male; mean age at T1 = 11.2, SD = 0.5). The
proportion of missing responses varied between 28.0% and 46.2%
by wave and cohort, whereas the number of respondents ranged
from 7,786 (T4, prepandemic) to 10,584 (T1, during pandemic). The
EQUITABLE SHIFTS IN YOUTH RESILIENCE? 3
greatest difference between cohorts was in T2 (during-pandemic
missing data: 46.2%, prepandemic: 31.9%). Overall, the missing
data patterns were similar across cohorts (for detailed analysis,
see Supplemental Tables 1–4).
Measures
Demographic factors included age, gender (female/male), parental
education as the highest education level of guardians, non-English
background (yes/no), and Aboriginal/Torres Strait Islander (yes/no).
These factors were extracted from the school enrollment records
and were parent reported. School-level factors included school
socioeconomic status index (Index of Educational Disadvantage)
and geographical remoteness index (Accessibility/Remoteness Index
of Australia).
ASE and CR were measured at four waves using a 5-point scale
varying from 0 (strongly disagree) to 4 (strongly agree). ASE,
measured with three items (e.g., “I am certain I can learn the skills
taught in school this year.” “Even if the work in school is hard, I can
learn it.” “If I have enough time, I can do a good job on all my school
work.”), had a Cronbach’s α ranging from .82 to .87 across time and
cohort. The measure has been validated as part of the Middle Years
Development Instrument (Schonert-Reichl et al., 2013). CR, also
measured with three items (“When I’m worried about something, I
make myself think about it in a different way that helps me feel
better.” “When I want to feel happier about something, I change
the way I’m thinking about it.” “When I want to feel less bad [e.g.,
sad, angry or worried], I change the way that I’m thinking about it.”),
had a Cronbach’s α ranging from .84 to .91 across waves and
cohorts. The measure was originally adopted from the Emotion
Regulation Questionnaire for Children andAdolescents and thereafter
shortened based on psychometric testing (Gregory & Brinkman,
2020; Gullone & Taffe, 2012).
Anxiety symptomswere measured at baseline using the “four-item
Worries scale” (e.g., “I worry a lot about mistakes that I make” and “I
worry a lot about things at school”; Gregory et al., 2019). Items were
rated on a scale from 0 (strongly disagree) to 4 (strongly agree).
Cronbach’s α was .83 in both cohorts. Peer belonging was measured
at baseline using a three-item scale (e.g., “I feel that I usually fit
in with other kids around me.”). Items were rated on a scale from
0 (strongly disagree) to 4 (strongly agree). Cronbach’s α was 85 in
both cohorts. Teacher support was measured at baseline using a
five-item scale (e.g., “Most of the teachers are interested in my well-
being.” “If I need extra help, I will receive it from my teacher.”).
Items were rated on a scale from 0 (strongly disagree) to 3 (strongly
agree). Cronbach’s α for this scale was .86 in both cohorts. For a
full list of items in all scales, see Supplemental Table 5.
Analytic Plan
The study analytic plan was preregistered, and all analytic
R-code, together with synthetic data and preregistration, is available
at https://doi.org/10.17605/OSF.IO/HKQ39. We report how we
determined our sample size, all data exclusions, all manipulations,
and all measures in the study. Survey responses were combined
with school enrolment data to include demographic factors, utilizing
the most recent available records. Further, survey responses from
both cohorts were merged to create a continuous developmental
trajectory spanning Grades 6 through 9 for both outcomes, using
grade as time. Cohort (prepandemic/during pandemic) was used as a
binary grouping variable. Full information maximum likelihood
estimation was employed for managing missing data during growth
modeling. Respondents with missing values on all outcomes in all
waves were excluded (301 respondents for ASE and 144 for CR).
To ensure the comparability of the cohorts, an examination of
all measures was conducted at baseline (Grade 6). This step was
essential as the interpretation of the results relied on the assumption
of baseline similarity between the cohorts. Moreover, intercorrela-
tions among key variables at baseline were explored.
To ensure the reliability and validity of the outcomes, we
formed latent measurement models and assessed their adequacy and
measurement invariance across time and cohorts using a series of
confirmatory factor analyses. Because we aimed to explore mean-
level changes, we anticipated achieving at least scalar measurement
invariance level (for details, see Supplemental Material).
Before modeling, we tabulated and plotted the mean trajectories
for both cohorts. Subsequently, we specified a series of unconditional
second-order latent growth curvemodels (LGCMs) for each outcome
and cohort individually. These LGCMs allowed us to scrutinize the
observed mean-level growth trajectories. We tested for no growth,
linear growth, month-specific linear growth, and quadratic growth
patterns. The month-specific growth pattern accounted for variations
based on the specific month of annual data collection. To decide on
the most appropriate shape for our growth models, we conducted
model fit comparisons. We evaluated the models based on fit indices
including and in priority order of (a) the Akaike information criterion
(AIC) and the Bayesian information criterion (BIC) where values
closer to 0 indicate a better fit, (b) comparative fit index (CFI) where
greater values indicate a better fit, and (c) the root-mean-square error
of approximation (RMSEA) where lower values indicate better fit
(Hu & Bentler, 1999). Across all growth models, the intercepts of
latent constructs were constrained to zero.
To assess possible pandemic effects, we specified conditional
multigroup LGCMs, separately for both outcomes, with cohort as a
grouping variable and gender, parental education level, and non-
English background as covariates (see Figure 1). To test cohort
differences in slope means, slope variances, and intercept–slope
covariances, we began with a constrained model where the growth
estimates were set equal across groups. Thereafter, we progressed by
releasing these equality constraints, permitting us to assess group
differences in the growth estimates by comparing the model fit to the
previous, more constrained model. Slope means indicate the average
rate of change over time within each group. Slope variances reflect
the variability in individual rates of change. Intercept–slope covariances
reflect the relationship between the initial value (intercept) and the
rate of change (slope). To evaluate differences between nested models,
we relied on ΔAIC values greater than 10 as evidence of a difference
and avoided using the likelihood ratio test and standardized root-mean-
square residual due to their sensitivity to sample size (Burnham &
Anderson, 2004).
To estimate a potential pandemic effect on a 4-year change in the
outcomes, we compared the average slope estimates between the
cohorts. Furthermore, we explored whether the pandemic widened
the interindividual differences in development by comparing slope
variances and intercept–slope covariances across cohorts. For
illustrative purposes, we categorized the samples into three groups
based on initial outcome levels: −0.5SE below the mean, at the
mean, and +0.5 SE above the mean.
4 REPO, HERKAMA, AND SALMIVALLI
Further analyzing the specified LGCMs, we examined the
associations between demographic covariates and outcome slope
means. We tested cohort differences in these associations by applying
equality constraints. Finally, we added anxiety, peer belonging, and
teacher support into the models. We examined how these socio-
emotional factors were associated with outcome development and
whether these associations differed between the cohorts.
Models were estimated using a robust maximum likelihood
estimator and adjusting standard errors for the nonindependence of
observations due to a hierarchical data structure (students nested in
schools). In all models, continuous predictors were centered at the
sample mean.
Results
Descriptive Results
The means and standard deviations for the demographic variables
and baseline values, along with cohort comparisons, are shown in
Table 1. According to t tests, the cohort differences in demographic
factors were nonexisting except in parental education level, where
the prepandemic cohort had somewhat lower average compared
with the during-pandemic cohort, M (SD) = 1.30 (0.88) and 1.34
(0.85), t = −3.22, d = −.04. The prepandemic cohort scored higher
in teacher support,M (SD) = 0.89 (0.60) and 0.81 (0.55), t = 10.20,
d = .15, yet lower in peer belonging,M (SD) = 1.08 (1.02) and 1.15
(0.93), t = −4.80, d = −0.07, compared with the during-pandemic
cohort. Regarding the outcomes, no baseline cohort differences were
identified. Supplemental Table 6 provides correlations among key
variables. The highest correlations were found between ASE and
teacher support (r = .61) and between ASE and CR (r = .47).
What Are the Typical Developmental Patterns Observed
in ASE and CR?
There was strong measurement invariance across time and
cohorts for ASE and CR (see Supplemental Table 7). Examination
of the mean-level changes in the prepandemic cohort suggested
declining trends in both outcomes (see Table 2). From Grades 6 to 9,
Figure 1
The Conceptual Model for Latent Growth Curve Models
Note. T = time.
Table 1
Descriptive Statistics for Demographics and Baseline Measurements by Cohort
Variable Range NA%
Prepandemic During pandemic
t dM SD M SD
Gender (1 = female) 0–1 0.00 0.49 0.50 0.49 0.50 0.42 0.01
Age T1 10–17 89.29 11.23 0.46 11.23 0.44 0.04 0.00
Aboriginal (1 = yes) 0–1 0.00 0.05 0.23 0.06 0.24 −1.49 −0.02
Non-English background (1 = yes) 0–1 0.00 0.24 0.43 0.25 0.43 −1.10 −0.01
Parental education level 0–3 10.92 1.30 0.88 1.34 0.85 −3.22 −0.04
School socioeconomic index (IOED) 1–7 0.00 4.47 1.74 4.51 1.74 −2.17 −0.03
School area (ARIA) 0–4 0.00 0.52 0.89 0.49 0.86 3.27 0.04
Academic self-efficacy T1 0–4 31.05 3.06 0.90 3.07 0.80 −0.39 −0.01
Cognitive reappraisal T1 0–4 30.48 2.49 0.92 2.50 0.90 −0.43 −0.01
Anxiety T1 0–4 29.95 1.93 1.13 1.91 1.04 1.13 0.02
Peer belonging T1 0–4 30.49 2.92 1.02 2.85 0.93 4.80 0.07
Teacher support T1 0–3 31.08 2.11 0.60 2.19 0.55 −10.20 −0.15
Note. Prepandemic N = 13,372, during pandemic N = 14,935. T1 = first wave. NA% = Proportion of missing values; IOED = Index of Educational
Disadvantage; ARIA = Accessibility/Remoteness Index of Australia.
EQUITABLE SHIFTS IN YOUTH RESILIENCE? 5
mean scores in ASE decreased from 3.06 (SD = 0.90) to 2.73 (SD =
0.84) and in CR from 2.49 (SD = 0.92) to 2.24 (SD = 0.93).
Did the Pandemic Affect the Development of
ASE and CR?
As shown in Table 2, mean scores for the during-pandemic cohort
indicated similar declines in both outcomes. By the last measure-
ment, the during-pandemic cohort showed slightly lower mean
levels and higher standard deviations in both ASE and CR compared
with the prepandemic cohort. However, the mean difference
between cohorts in either outcome did not increase after the second
measurement (Grade 7). Mean scores are plotted in Supplemental
Figure 2.
Next, we fitted unconditional second-order LGCMs, separately
for both outcomes and cohorts. After testing different shapes
for mean-level growth and comparing model fits, we chose the
month-specific linear growth as the best fitting pattern. These
models demonstrated a good fit to data across both outcomes and
cohorts with all models showing CFI > .98 and RMSEA < .03 (see
Supplemental Table 8).
The results from best fit unconditional multigroup LGCMs
indicated negative estimated slope means across all conditions,
indicating a decrease in both outcomes for both cohorts (see Table 3).
Results from the conditional multigroup LGCMs, also shown in
Table 3, similarly indicated negative slope means across all
conditions. With ASE, there was no clear cohort difference in the
average slopes (ß = −0.11, SE = 0.008, ß = −0.11, SE = 0.006
respectively, difference in standardized ß = 0.16). Model fit
comparisons with invariance constraints showed no evidence for
cohort difference (ΔAIC = −7 ΔBIC = 2; see Supplemental
Table 9). For CR, the during-pandemic cohort showed a slightly
steeper decline compared with the prepandemic cohort (ß = −0.06,
SE = 0.010, ß = −0.09, SE = 0.006, difference in standardized ß =
0.30), and model fit analyses indicated difference between cohorts
(ΔAIC = −14, ΔBIC = −5).
Did the Pandemic Exacerbate Between-Person
Differences in the Development of ASE and CR?
For ASE, model fit analysis indicated a cohort difference in slope
variance (ΔAIC = −12, ΔBIC = −4), and as shown in Table 3,
the variance was smaller in the during-pandemic cohort (ß = 0.02,
SE = 0.002) compared with the prepandemic cohort (ß = 0.04,
SE = 0.003). As illustrated in Figure 2, the slope distribution in
the during-pandemic cohort is more centered around the mean,
indicating smaller between-person differences in the rate of change.
For CR, the model fit analyses indicated no cohort differences
in slope variance, yet the estimates and graphical examination
indicated a similar pattern as with ASE (see Figure 2).
For intercept–slope covariance, the results indicated a similar
pattern for both outcomes. Model fit analyses indicated clear cohort
differences for ASE (ΔAIC = −79, ΔBIC = −71) and CR (ΔAIC =
−38, ΔBIC = −30). The covariances were smaller in the during-
pandemic cohort (ASE: ß = −0.02, SE = 0.005, CR: ß = −0.03,
SE = 0.006) compared with the prepandemic cohort (ASE: ß =
−0.09, SE = 0.008, CR: ß = −0.09, SE = 0.009).
This difference is illustrated in Figure 3. In the prepandemic
cohort, those starting low (in red) showed no decline in time,
catching others in growth. In the during-pandemic cohort, this
compensation effect was not present, as also the ones starting low
showed a declining trend. In the during-pandemic cohort, the
initially high-scoring individuals (in green) showed more favorable
development than their prepandemic counterparts, but only after a
steep decline during the first wave of the pandemic (Grade 7, year
2020). In sum, this led to greater interindividual differences at
the final wave (Grade 9) in the during-pandemic cohort.
What Factors Predicted the Development in ASE and CR
and Did the Pandemic Make a Difference in These?
After finding substantial interindividual variation in the slopes for
both outcomes, we proceeded to identify predictors for the slopes.
The regression estimates for the covariates are shown in Table 3. A
negative estimate implies that higher values of the predictor are
associated with steeper declines in the outcome over time. Overall,
gender had a negative association, whereas non-English background
and parental education had zero to positive associations with the
slopes, in both outcomes and in both cohorts. These indicate that
students with male gender, non-English background, and more
highly educated parents had somewhat less decrease in ASE and CR
from Grades 6 to 9 compared with their counterparts. For these
associations, the model fit comparison analysis revealed no cohort
differences (see Supplemental Table 10).
Table 2
Descriptives of Outcomes by Time and Cohort
Variable
Prepandemic During pandemic
t dN NA% M SD N NA% M SD
Academic self-efficacy
T1—Grade 6 9,475 29.14 3.06 0.90 10,044 32.75 3.07 0.80 −0.39 −0.01
T2—Grade 7 8,961 32.99 3.07 0.89 7,639 48.85 2.92 0.82 11.06 0.17
T3—Grade 8 8,835 33.93 2.78 0.84 9,885 33.81 2.80 0.81 −1.72 −0.03
T4—Grade 9 7,526 43.72 2.73 0.84 8,159 45.37 2.64 0.87 6.53 0.10
Cognitive reappraisal
T1—Grade 6 9,522 28.79 2.49 0.92 10,156 32.00 2.50 0.90 −0.43 −0.01
T2—Grade 7 9,037 32.42 2.35 0.93 7,769 47.98 2.25 0.95 6.79 0.11
T3—Grade 8 8,958 33.01 2.27 0.91 10,019 32.92 2.16 0.93 7.67 0.11
T4—Grade 9 7,606 43.12 2.24 0.93 8,316 44.32 2.14 0.95 6.87 0.11
Note. Response scales 0–4. Results based on mean scales. NA = missing values; T = time.
6 REPO, HERKAMA, AND SALMIVALLI
Regarding the socioemotional predictors, for anxiety symptoms
and peer belonging, model fit comparison indicated no cohort
differences (see Supplemental Table 11). For teacher support, the
analysis indicated a cohort difference in CR (ΔAIC = −11, ΔBIC =
−10), but not in ASE (ΔAIC = 3, ΔBIC = 4). The association was
negative and stronger in the during-pandemic cohort (ß = 0.11,
SE = 0.010) compared with the prepandemic cohort (ß = 0.07,
SE = 0.009). This indicates that those with higher initial levels
of teacher support had more negative development in CR and
especially in the during-pandemic cohort. This interaction is
illustrated in the Supplemental Figure 1.
Discussion
The present longitudinal multicohort study aimed to investigate
the long-term effects of the COVID-19 pandemic on ASE and CR
beliefs among early adolescents in South Australia. It tracked two
cohorts over 4 years: one that experienced their early adolescence
(11–14 years) before the pandemic (2016–2019) and another during
the pandemic (2019–2022). The study examined the normative
development of ASE and CR, their alteration during the pandemic,
the impact of demographic and socioemotional factors, as well as the
pandemic’s possible role in magnifying individual differences.
Did the Pandemic Affect ASE and CR Development?
The prepandemic cohort showed a stable yet gradually declining
trend in both ASE and CR. Given the prior mixed findings on
normative developmental trends, this finding is notable and aligns
with several studies among the same age group (Gregory &
Brinkman, 2020; Gullone et al., 2010; Vestad et al., 2022;
Zimmermann & Iwanski, 2014). The during-pandemic cohort also
showed developmental declines in both domains. However, the
cohort comparison indicated that these decreases largely echoed
normative developmental changes: The 4-year ASE trajectory was
similar to that of the prepandemic cohort, with a slightly deeper
decline in CR. This suggests that pandemic’s long-term impact was
marginally small yet negative on CR and almost negligible on ASE.
This aligns with one previous study, which found that even the
ones who perceived a negative effect had no observable changes in
ASE (Talsma et al., 2021).
Table 3
Cohort Comparison on Latent Growth and Slope Covariant Estimates
Parameter
Prepandemic During pandemic
Std. B.DiffB (CIs) SE Std. B B (CIs) SE Std. B
Academic self-efficacy, unconditional model (CFI = 0.981, AIC = 485,049, BIC = 485,428, RMSEA = 0.032)
Intercept 3.13 [3.09, 3.16] 0.019 3.01 [2.98, 3.04] 0.014 0.19a
Variance (intercept) 0.53 [0.48, 0.57] 0.024 0.29 [0.26, 0.32] 0.014 0.37b
Slope −0.13 [−0.15, −0.12] 0.006 −0.63 −0.13 [−0.13, −0.12] 0.005 −0.79 0.16b
Variance (slope) 0.05 [0.04, 0.05] 0.003 1.00 0.03 [0.02, 0.03] 0.002 1.00 0
Covariance (intercept, slope) −0.09 [−0.11, −0.08] 0.008 −0.61 −0.02 [−0.03, −0.01] 0.005 −0.27 −0.34
Academic self-efficacy, Model 1 (CFI = 0.977, AIC = 476,348, BIC = 476,826, RMSEA = 0.030)
Slope −0.11 [−0.13, −0.09] 0.008 −0.51 −0.11 [−0.12, −0.09] 0.006 −0.67 0.16
Variance (slope) 0.04 [0.04, 0.05] 0.003 0.98 0.02 [0.02, 0.03] 0.002 0.95 0.03
Covariance (intercept, slope) −0.09 [−0.11, −0.08] 0.008 −0.61 −0.02 [−0.03, −0.01] 0.005 −0.28 −0.33
Sex ∼S −0.06 [−0.07, −0.04] 0.007 −0.14 −0.06 [−0.07, −0.04] 0.007 −0.17 0.03
Non-English background ∼S 0.01 [−0.00, 0.03] 0.009 0.03 0.03 [0.01, 0.04] 0.008 0.08 −0.05
Parental education level ∼S 0.01 [−0.00, 0.02] 0.005 0.03 0.02 [0.01, 0.02] 0.004 0.09 −0.06
Academic self-efficacy, Model 2 (CFI = 0.992, AIC = 485,579, BIC = 486,058, RMSEA = 0.019)
Anxiety ∼S −0.05 [−0.06, −0.04] 0.005 −0.22 −0.04 [−0.05, −0.03] 0.005 −0.21 −0.01
Peer belonging ∼S −0.05 [−0.06, −0.04] 0.005 −0.22 −0.04 [−0.05, −0.03] 0.005 −0.21 −0.01
Teacher support ∼S −0.15 [−0.17, −0.13] 0.01 −0.38 −0.15 [−0.16, −0.13] 0.009 −0.42 0.04
Cognitive reappraisal, unconditional model (CFI = 0.993, AIC = 494,430, BIC = 494,810, RMSEA = 0.021)
Intercept 2.48 [2.45, 2.51] 0.013 2.41 [2.38, 2.43] 0.013 0.11a
Variance (intercept) 0.51 [0.46, 0.57] 0.027 0.39 [0.36, 0.43] 0.017 0.18b
Slope −0.09 [−0.11, −0.08] 0.007 −0.40 −0.11 [−0.12, −0.11] 0.004 −0.69 0.29b
Variance (slope) 0.05 [0.05, 0.06] 0.004 1.00 0.03 [0.02, 0.03] 0.003 1.00 0
Covariance (intercept, slope) −0.10 [−0.11, −0.08] 0.009 −0.57 −0.03 [−0.04, −0.02] 0.006 −0.29 −0.28
Cognitive reappraisal, Model 1 (CFI = 0.971, AIC = 406,681, BIC = 407,258, RMSEA = 0.031)
Slope −0.06 [−0.08, −0.04] 0.01 −0.27 −0.09 [−0.11, −0.08] 0.006 −0.57 0.3
Variance (slope) 0.05 [0.04, 0.06] 0.004 0.93 0.02 [0.02, 0.03] 0.003 0.92 0.01
Covariance (intercept, slope) −0.09 [−0.11, −0.07] 0.009 −0.58 −0.03 [−0.04, −0.02] 0.006 −0.32 −0.26
Sex ∼S −0.12 [−0.13, −0.10] 0.009 −0.25 −0.08 [−0.10, −0.07] 0.007 −0.25 0
Non-English background ∼S 0.05 [0.03, 0.07] 0.012 0.09 0.04 [0.03, 0.06] 0.008 0.12 −0.03
Parental education level ∼S 0.01 [0.00, 0.02] 0.005 0.04 0.01 [−0.00, 0.01] 0.004 0.03 0.01
Cognitive reappraisal, Model 2 (CFI = 0.991, AIC = 419,213, BIC = 419,790, RMSEA = 0.018)
Anxiety ∼S 0.01 [−0.00, 0.02] 0.005 0.03 0.01 [0.01, 0.02] 0.004 0.08 −0.05
Peer belonging ∼S −0.04 [−0.05, −0.03] 0.005 −0.17 −0.03 [−0.04, −0.02] 0.005 −0.16 −0.01
Teacher support ∼S −0.07 [−0.09, −0.06] 0.009 −0.19 −0.11 [−0.13, −0.09] 0.01 −0.32 0.13
Note. Std. B.Diff. = cohort difference in standardized coefficients; B = unstandardized coefficients; SE = standard error; CIs = confidence intervals; CFI =
comparative fit index; AIC = Akaike information criterion; BIC = Bayesian information criterion; RMSEA = root-mean-square error of approximation.
a Cohen’s d. b Standardized difference.
EQUITABLE SHIFTS IN YOUTH RESILIENCE? 7
To analyze whether the pandemic exacerbated between-person
differences, we categorized students based on their initial outcome
levels. The results indicated that in the prepandemic cohort, CR and
ASE scores converged over time; students with initially low scores
displayed positive growth, in contrast to their peers who experienced
a negative decline, aligning with an earlier prepandemic study (Herd
et al., 2020). However, the during-pandemic cohort did not exhibit
this compensation effect, with those having low initial scores also
declining over time. Additionally, high-starting students showed a
less steep decline compared with their prepandemic counterparts.
These findings were consistent with the observation of more
uniform slope variance within the during-pandemic cohort.
In essence, individuals with low initial scores seemed somewhat
negatively impacted by the pandemic, whereas those with higher
initial scores showed negligible or modest positive effects in both
outcomes. This pattern suggests a modest “buffering effect” against
adverse long-term pandemic effects in both outcomes. Consequently,
the during-pandemic cohort exhibited greater between-person
differences by the final measurement, highlighting a pronounced,
albeit modest, exacerbating pandemic effect on both ASE and CR.
What Factors Influenced Development, and Did the
Pandemic Alter These Effects?
When examining only the during-pandemic cohort, our results
initially mirrored existing pandemic studies, showing that female
gender correlated with less favorable development (Madigan et al.,
2023). However, when considering the normative development
based on the prepandemic cohort, we find no evidence for a
disproportionate pandemic effect. The same applied for parental
education.
Unsurprisingly, our results indicate that students with preexisting
anxiety symptoms had lower levels of CR. Intriguingly, baseline
anxiety did not lead to greater decline in CR during adolescence, after
Figure 3
Comparative Analysis of Outcome Trajectories Based on Initial Levels Across Cohorts
Academic Self−Efficacy Cognitive Reappraisal
6 7 8 9 6 7 8 9
1.5
2.0
2.5
3.0
3.5
Grade Level
Pr
ed
ict
ed
 V
a
lu
es
Pre−pandemic During−pandemic Intercept −0.5 SD Mean +0.5 SD
Figure 2
Distributions of Predicted Slope Values by Cohort
0
2
4
6
−0.75 −0.50 −0.25 0.00 0.25
Slope mean
D
en
si
ty
Pre−pandemic During−pandemic
Academic Self−Efficacy
0
2
4
−0.50 −0.25 0.00 0.25
Slope mean
Pre−pandemic During−pandemic
Cognitive Reappraisal
Note. Estimated slopemean represents overall change in 4 years. The scale for both outcomeswas 0–4.
8 REPO, HERKAMA, AND SALMIVALLI
adjusting for factors like peer belonging and teacher support. Nor was
it linked to any pandemic effect. This aligns with previous studies
where preexisting anxiety has shown both positive and negative
associations with adverse pandemic effects (Bouter et al., 2023;
Haikalis et al., 2022; Morales et al., 2022; Skinner et al., 2023).
Correspondingly, students with lower peer belonging scores
exhibited reduced ASE and CR compared with their peers, yet this
factor was unrelated to pandemic-related changes. Interestingly, these
students tended to catch up in ASE and CR growth, similarly in both
cohorts. While prior research has focused on the immediate negative
associations of loneliness, a few have noticed that adolescents with
preexisting loneliness have not beenmore but often less affected by the
pandemic (Hamza et al., 2021; Mlawer et al., 2022; Repo et al., 2023).
Finally, students with low initial teacher support typically
displayed lower ASE and ER, yet less pandemic impact. Compared
with their peers, they showed catch-up in growth, particularly in ER,
and somewhat more during the pandemic. It is possible that students
with stronger teacher relationships perceived a greater loss when
schools closed, while those used to less support experienced less
disruption. Intriguingly, those with weaker initial teacher support
appeared more resilient in facing pandemic conditions.
In summary, while various vulnerabilities were associated with
lower ASE and ER, only gender showed an association over
time, independent of the pandemic impact. Notably, none of the
vulnerabilities predicted a negative pandemic effect. The findings
suggest that students’ adaptability and resource maintenance,
despite initial vulnerabilities, were key in mitigating the adverse
effects on ASE and ER, indicating resilience beyond the expected
negative outcomes of these vulnerabilities.
Limitations
In interpreting the results of this study, several limitations must be
considered. First, the constructs measured with a limited number of
items, despite achieving good internal consistency, represent only
specific aspects of the outcomes. Second, using only adolescents’
self-reports may have led to overestimated effects due to shared
method variance, yet this would not significantly affect the main
conclusions. Third, variation in data collection timing during the
pandemic’s initial year, with some data gathered before and some
during the first pandemic wave, introduces uncertainty for that
period’s results. Had all 2020 data been collected after the pandemic
began, the results might show a steeper decline that year. However,
this does not significantly affect the overall findings, as the study
focused on the 4-year longitudinal change. The fourth limitation is
potential selection bias, given the attrition rates across waves. Those
who did not participate might have faced greater challenges, as our
missing data analysis suggested. However, missing data patterns
did not differ between cohorts. Nonetheless, caution is advised in
generalizing our findings. Finally, the study was limited to South
Australia, with its distinct pandemic experiences, policies, and
cultural factors. Therefore, the extent of pandemic effects might vary
in other cultural contexts. Nonetheless, our main findings could offer
valuable perspectives for future research in diverse cultural settings.
Conclusion
As its main conclusion, this study highlights the widespread
methodological shortcomings in pandemic research (see also
Repo, 2024). Many effects often attributed to the pandemic may
actually be part of normative developmental patterns. Furthermore,
self-perceived pandemic effects may be unrelated to actual long-term
effects. The longitudinal cohort comparison was crucial to discern
the actual extent of average pandemic effects, to pinpoint both
exacerbation and buffering effects, and to discover that none of the
preexisting vulnerabilities examined predicted adverse pandemic
effects. This underscores the need for caution in interpreting all
pandemic studies that fail to account for developmental changes.
Such caution is also vital when researching future crises and other
widespread phenomena. Overall, our research emphasizes the critical
importance of understanding normative youth development and the
need for robust longitudinal data infrastructures, as exemplified by
the Wellbeing and Engagement Collection in South Australia.
Furthermore, the study predominantly illustrates the inherent
resilience in youth. Despite facing numerous pandemic-related
disruptions and negative emotional impacts during a critical
developmental phase, early adolescents largely maintained similar
trajectories in ASE and CR as their prepandemic counterparts. The
results indicate stability in self-efficacy beliefs despite significant
contextual and emotional shifts.
In addition, the study emphasizes the diversity of individual
experiences. It indicated that the pandemic led to a moderate
exacerbation of between-person differences, particularly affecting
those with initially low in ASE and CR. However, the findings
advise against a simplistic view that individuals with prepandemic
vulnerabilities faced the worst effects. Resilience research shows
that past adversities can sometimes buffer future stress impacts—yet
opportunities for stress-induced resilience are not equally available for
all (Rutter, 2012). This highlights the need for a more nuanced
approach to the sociopsychological aspects of resilience development.
Finally, while this study focused on the development of ASE and
CR among South Australian adolescents, its implications reach
beyond these specific domains. In a broader sense, the findings
emphasize the crucial need for robust data and designs to account for
normative developmental changes when interpreting any temporal
changes observed among developing adolescents.
References
Alamolhoda, S. H., Zare, E., HakimZadeh, A., Zalpour, A., Vakili, F.,
Chermahini, R. M., Ebadifard, R., Masoumi, M., Khaleghi, N., & Nasiri,
M. (2022). Adolescent mental health during covid-19 pandemics: A
systematic review. International Journal of Adolescent Medicine and
Health, 35(1), 41–60. https://doi.org/10.1515/ijamh-2022-0058
Aldao, A., Nolen-Hoeksema, S., & Schweizer, S. (2010). Emotion-
regulation strategies across psychopathology: A meta-analytic review.
Clinical Psychology Review, 30(2), 217–237. https://doi.org/10.1016/j.cpr
.2009.11.004
Alessandri, G., Vecchione, M., & Caprara, G. V. (2015). Assessment
of regulatory emotional self-efficacy beliefs: A review of the status of
the art and some suggestions to move the field forward. Journal of
Psychoeducational Assessment, 33(1), 24–32. https://doi.org/10.1177/
0734282914550382
Australian Bureau of Statistics. (2022). Australian schools, 2021. https://
www.abs.gov.au/statistics/people/education/schools/2021
Bagheri Sheykhangafshe, F., Hajialiani, V., & Hasani, J. (2022). The role of
emotion regulation strategies in mental health during the COVID-19
pandemic: A systematic review. Journal of Modern Medical Information
Sciences, 8(2), 196–207. https://doi.org/10.32598/JRH.11.6.1922.1
EQUITABLE SHIFTS IN YOUTH RESILIENCE? 9
Bandura, A. (1997). Self-efficacy: The exercise of control. WH Freeman/
Times Books/Henry Holt.
Beames, J. R., Li, S. H., Newby, J. M., Maston, K., Christensen, H., &
Werner-Seidler, A. (2021). The upside: Coping and psychological
resilience in Australian adolescents during the COVID-19 pandemic.
Child and Adolescent Psychiatry and Mental Health, 15(1), Article 77.
https://doi.org/10.1186/s13034-021-00432-z
Bouter, D. C., Zarchev, M., de Neve-Enthoven, N. G. M., Ravensbergen,
S. J., Kamperman, A. M., Hoogendijk, W. J. G., & Grootendorst-van Mil,
N. H. (2023). A longitudinal study of mental health in at-risk adolescents
before and during the COVID-19 pandemic. European Child &
Adolescent Psychiatry, 32, 1109–1117. https://doi.org/10.1007/s00787-
021-01935-y
Burnham, K. P., & Anderson, D. R. (Eds.). (2004). Model selection and
multimodel inference. Springer. https://doi.org/10.1007/b97636
Domitrovich, C. E., Durlak, J. A., Staley, K. C., & Weissberg, R. P. (2017).
Social–emotional competence: An essential factor for promoting positive
adjustment and reducing risk in school children. Child Development,
88(2), 408–416. https://doi.org/10.1111/cdev.12739
Durlak, J. A., Weissberg, R. P., Dymnicki, A. B., Taylor, R. D., &
Schellinger, K. B. (2011). The impact of enhancing students’ social and
emotional learning: A meta-analysis of school-based universal interven-
tions.Child Development, 82(1), 405–432. https://doi.org/10.1111/j.1467-
8624.2010.01564.x
Edwards, B., Barnes, R., Rehill, P., Ellen, L., Zhong, F., Killigrew, A.,
Gonzalez, R., Sheard, E., Zhu, R., & Philips, T. (2022). Variation in policy
response to COVID-19 across Australian states and territories. Blavatnik
School of Government, University of Oxford.
Fong, C. J., Berger, C., & Bailey, T. (2022). Academic motivation in a
pandemic context: A conceptual review of prominent theories and an
integrative model. Educational Psychologist, 57(3), 153–170. https://
doi.org/10.1080/00461520.2022.2062100
Gorman, D. M. (2023). COVID-19 publications in top-ranked public health
journals during the first phase of the pandemic. Quantitative Science
Studies, 4(2), 535–546. https://doi.org/10.1162/qss_a_00257
Gregory, T., & Brinkman, S. (2020).Wellbeing and Engagement Collection
(WEC): History of the WEC in the South Australian school system and
psychometric properties of the WEC survey instrument. Fraser Mustard
Centre. https://www.education.sa.gov.au/docs/system-performance/frase
r-mustard-centre/sa-wec-technical-report.pdf
Gregory, T., Engelhardt, D., Lewkowicz, A., Luddy, S., Guhn, M.,
Gadermann, A., Schonert-Reichl, K., & Brinkman, S. (2019). Validity of
the middle years development instrument for population monitoring of
student wellbeing in australian school children.Child Indicators Research,
12(3), 873–899. https://doi.org/10.1007/s12187-018-9562-3
Gregory, T., Lewkowicz, A., Engelhardt, D., Stringer, A., Luddy, S., &
Brinkman, S. A. (2022). Data resource profile: The South Australian well-
being and engagement collection (WEC). International Journal of
Epidemiology, 51(1), 16–16g. https://doi.org/10.1093/ije/dyab103
Gross, J. J., & Thompson, R. A. (2007). Emotion regulation: Conceptual
foundations. Guilford Press.
Gullone, E., Hughes, E. K., King, N. J., & Tonge, B. (2010). The normative
development of emotion regulation strategy use in children and adolescents:
A 2-year follow-up study. Journal of Child Psychology and Psychiatry, and
Allied Disciplines, 51(5), 567–574. https://doi.org/10.1111/j.1469-7610
.2009.02183.x
Gullone, E., & Taffe, J. (2012). The Emotion Regulation Questionnaire for
Children and Adolescents (ERQ-CA): A psychometric evaluation. Psycho-
logical Assessment, 24(2), 409–417. https://doi.org/10.1037/a0025777
Haikalis, M., Doucette, H., Meisel, M. K., Birch, K., & Barnett, N. P. (2022).
Changes in college student anxiety and depression from pre- to during-
COVID-19: Perceived stress, academic challenges, loneliness, and
positive perceptions. Emerging Adulthood, 10(2), 534–545. https://
doi.org/10.1177/21676968211058516
Hale, T., Webster, S., Petherick, A., Phillips, T., & Kira, B. (2020). Oxford
COVID-19 government response tracker (OxCGRT). Blavatnik School of
Government, University of Oxford. https://www.bsg.ox.ac.uk/research/co
vid-19-government-response-tracker
Hamza, C. A., Ewing, L., Heath, N. L., & Goldstein, A. L. (2021). When
social isolation is nothing new: A longitudinal study on psychological
distress during COVID-19 among university students with and without
preexisting mental health concerns. Canadian Psychology/Psychologie
Canadienne, 62(1), 20–30. https://doi.org/10.1037/cap0000255
Hayes, S., Carlyle, M., Haslam, S. A., Haslam, C., & Dingle, G. (2022).
Exploring links between social identity, emotion regulation, and loneliness
in those with and without a history of mental illness. British Journal of
Clinical Psychology, 61(3), 701–734. https://doi.org/10.1111/bjc.12358
Herd, T., King-Casas, B., &Kim-Spoon, J. (2020). Developmental changes in
emotion regulation during adolescence: Associations with socioeconomic
risk and family emotional context. Journal of Youth and Adolescence,
49(7), 1545–1557. https://doi.org/10.1007/s10964-020-01193-2
Hofmann, S. G. (2014). Interpersonal emotion regulationmodel of mood and
anxiety disorders. Cognitive Therapy and Research, 38(5), 483–492.
https://doi.org/10.1007/s10608-014-9620-1
Honicke, T., & Broadbent, J. (2016). The influence of academic self-efficacy
on academic performance: A systematic review. Educational Research
Review, 17, 63–84. https://doi.org/10.1016/j.edurev.2015.11.002
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance
structure analysis: Conventional criteria versus new alternatives.
Structural Equation Modeling, 6(1), 1–155. https://doi.org/10.1080/
10705519909540118
Kauhanen, L., Wan Mohd Yunus, W. M. A., Lempinen, L., Peltonen, K.,
Gyllenberg, D., Mishina, K., Gilbert, S., Bastola, K., Brown, J. S. L., &
Sourander, A. (2023). A systematic review of the mental health changes of
children and young people before and during the COVID-19 pandemic.
European Child & Adolescent Psychiatry, 32, 995–1013. https://doi.org/
10.1007/s00787-022-02060-0
LeMoult, J., & Gotlib, I. H. (2019). Depression: A cognitive perspective.
Clinical Psychology Review, 69, 51–66. https://doi.org/10.1016/j.cpr
.2018.06.008
Liu, A. (2019). Can non-cognitive skills compensate for background
disadvantage?—The moderation of non-cognitive skills on family
socioeconomic status and achievement during early childhood and early
adolescence. Social Science Research, 83, Article 102306. https://doi.org/
10.1016/j.ssresearch.2019.04.019
Madigan, S., Racine, N., Vaillancourt, T., Korczak, D. J., Hewitt, J. M. A.,
Pador, P., Park, J. L., McArthur, B. A., Holy, C., & Neville, R. D. (2023).
Changes in depression and anxiety among children and adolescents
from before to during the COVID-19 pandemic: A systematic review and
meta-analysis. JAMA Pediatrics, 177(6), 567–581. https://doi.org/10
.1001/jamapediatrics.2023.0846
McRae, K., & Gross, J. J. (2020). Emotion regulation. Emotion, 20(1), 1–9.
https://doi.org/10.1037/emo0000703
Mlawer, F., Moore, C. C., Hubbard, J. A., & Meehan, Z. M. (2022). Pre-
pandemic peer relations predict adolescents’ internalizing response to
Covid-19. Research on Child and Adolescent Psychopathology, 50(5),
649–657. https://doi.org/10.1007/s10802-021-00882-1
Morales, S., Zeytinoglu, S., Lorenzo, N. E., Chronis-Tuscano, A., Degnan, K. A.,
Almas, A. N., Pine, D. S., & Fox, N. A. (2022). Which anxious adolescents
were most affected by the COVID-19 pandemic? Clinical Psychological
Science, 10(6), 1044–1059. https://doi.org/10.1177/21677026211059524
National Advisory Board on Research Ethics. (2009). Ethical principles of
research in the humanities and social and behavioural sciences and
proposals for ethical review. https://www.tenk.fi/sites/tenk.fi/files/ethica
lprinciples.pdf
Repo, J. (2024).Worst of times, best of times: Assessing pandemic impact on
adolescents [Doctoral dissertation, University of Turku]. https://urn.fi/
URN:ISBN:978-951-29-9997-2
10 REPO, HERKAMA, AND SALMIVALLI
Repo, J., Herkama, S., Yanagida, T., & Salmivalli, C. (2023). Transition to
emerging adulthood during the COVID-19 pandemic: Changes in anxiety
and the role of inclusion/exclusion experiences. European Journal of
Developmental Psychology, 20(4), 649–665. https://doi.org/10.1080/
17405629.2023.2183651
Repo, J., Smith, E., Kilpi-Jakonen, E., & Reimer, D. (in press). Student well-
being, social inequality and educational transitions before, during and after
the pandemic. Social Indicators Research.
Restad, F., &Mølstad, C. E. (2021). Social and emotional skills in curriculum
reform: A red line for measurability? Journal of Curriculum Studies, 53(4),
435–448. https://doi.org/10.1080/00220272.2020.1716391
Rutter, M. (2012). Resilience as a dynamic concept. Development and
Psychopathology, 24(2), 335–344. https://doi.org/10.1017/S095457
9412000028
Samji, H., Wu, J., Ladak, A., Vossen, C., Stewart, E., Dove, N., Long, D., &
Snell, G. (2022). Review: Mental health impacts of the COVID-19
pandemic on children and youth—A systematic review. Child and
Adolescent Mental Health, 27(2), 173–189. https://doi.org/10.1111/ca
mh.12501
Sanchis-Sanchis, A., Grau, M. D., Moliner, A.-R., & Morales-Murillo, C. P.
(2020). Effects of age and gender in emotion regulation of children and
adolescents. Frontiers in Psychology, 11, Article 946. https://doi.org/10
.3389/fpsyg.2020.00946
Schonert-Reichl, K. A., Guhn,M., Gadermann, A.M., Hymel, S., Sweiss, L.,
& Hertzman, C. (2013). Development and validation of the Middle Years
Development Instrument (MDI): Assessing children’s well-being and
assets across multiple contexts. Social Indicators Research, 114(2), 345–
369. https://doi.org/10.1007/s11205-012-0149-y
Shergold, P., Broadbent, J., & Marshall, I. (2022). Fault lines: An
independent review into australia’s response to COVID-19. John and
Myriam Wylie Foundation, Minderoo Foundation, and Paul Ramsay
Foundation. https://apo.org.au/node/320067
Skinner, A. T., De Luca, L., Nocentini, A., & Menesini, E. (2023). Co-
development of internalizing symptoms and regulatory emotional self-
efficacy in adolescence: Time-varying effects of COVID-19-related stress
and social support. International Journal of Behavioral Development,
47(5), 433–443. https://doi.org/10.1177/01650254231189743
Smithers, L. G., Sawyer, A. C. P., Chittleborough, C. R., Davies, N. M.,
Davey Smith, G., & Lynch, J. W. (2018). A systematic review and meta-
analysis of effects of early life non-cognitive skills on academic,
psychosocial, cognitive and health outcomes. Nature Human Behaviour,
2(11), 867–880. https://doi.org/10.1038/s41562-018-0461-x
Storen, R., & Corrigan, N. (2020). COVID-19: A chronology of state and
territory government announcements. https://www.aph.gov.au/About_Parlia
ment/Parliamentary_departments/Parliamentary_Library/pubs/rp/rp2021/Chro
nologies/COVID-19StateTerritoryGovernmentAnnouncements
Strasser, K., Arias, P., Alessandri, F., Turner, P., Villarroel, T., Aldunate,
C. P., & Montt, M. E. (2023). Adolescents’ academic self-efficacy and
emotions during the COVID-19 pandemic: A latent profile analysis of
family and school risk factors. School Psychology, 38(2), 88–99. https://
doi.org/10.1037/spq0000523
Strohmeier, D., & Branje, S. (2023). Human development in times of the
COVID-19 pandemic. European Journal of Developmental Psychology,
20(4), 581–596. https://doi.org/10.1080/17405629.2023.2211256
Tak, Y. R., Brunwasser, S. M., Lichtwarck-Aschoff, A., & Engels, R. C.
(2017). The prospective associations between self-efficacy and depressive
symptoms from early to middle adolescence: A cross-lagged model.
Journal of Youth and Adolescence, 46(4), 744–756. https://doi.org/10
.1007/s10964-016-0614-z
Talsma, K., Robertson, K., Thomas, C., & Norris, K. (2021). COVID-19
beliefs, self-efficacy and academic performance in first-year university
students: Cohort comparison and mediation analysis. Frontiers in
Psychology, 12, Article 643408. https://doi.org/10.3389/fpsyg.2021.643408
Talsma, K., Schüz, B., Schwarzer, R., & Norris, K. (2018). I believe,
therefore I achieve (and vice versa): A meta-analytic cross-lagged panel
analysis of self-efficacy and academic performance. Learning and
Individual Differences, 61, 136–150. https://doi.org/10.1016/j.lindif
.2017.11.015
Ungar, M., Ghazinour, M., & Richter, J. (2013). Annual research review:
What is resilience within the social ecology of human development?
Journal of Child Psychology and Psychiatry, and Allied Disciplines,
54(4), 348–366. https://doi.org/10.1111/jcpp.12025
Van Lissa, C. J., Brandmaier, A. M., Brinkman, L., Lamprecht, A.-L.,
Peikert, A., Struiksma, M. E., & Vreede, B. M. I. (2021). WORCS: A
workflow for open reproducible code in science. Data Science, 4(1), 29–
49. https://doi.org/10.3233/DS-210031
Vaughn, A. A., Roesch, S. C., & Aldridge, A. A. (2009). Stress-related
growth in racial/ethnic minority adolescents: Measurement structure and
validity. Educational and Psychological Measurement, 69(1), 131–145.
https://doi.org/10.1177/0013164408318775
Vestad, L., Bru, E., & Virtanen, T. E. (2022). Changes in academic efficacy
beliefs in the first year of lower secondary school. Is it related to changes in
social and emotional competencies? Educational Psychology, 42(7), 838–
856. https://doi.org/10.1080/01443410.2022.2093333
Willner, C. J., Hoffmann, J. D., Bailey, C. S., Harrison, A. P., Garcia, B., Ng,
Z. J., Cipriano, C., &Brackett, M. A. (2022). The development of cognitive
reappraisal from early childhood through adolescence: A systematic
review and methodological recommendations. Frontiers in Psychology,
13, Article 875964. https://doi.org/10.3389/fpsyg.2022.875964
Wright, N., Hill, J., Sharp, H., Refberg-Brown, M., Crook, D., Kehl, S., &
Pickles, A. (2024). COVID-19 pandemic impact on adolescent mental
health: A reassessment accounting for development. European Child &
Adolescent Psychiatry, 33(8), 2615–2627. https://doi.org/10.1007/s00787-
023-02337-y
Young, K. S., Sandman, C. F., & Craske,M. G. (2019). Positive and negative
emotion regulation in adolescence: Links to anxiety and depression. Brain
Sciences, 9(4), Article 76. https://doi.org/10.3390/brainsci9040076
Zimmermann, P., & Iwanski, A. (2014). Emotion regulation from
early adolescence to emerging adulthood and middle adulthood: Age
differences, gender differences, and emotion-specific developmental
variations. International Journal of Behavioral Development, 38(2), 182–
194. https://doi.org/10.1177/0165025413515405
Received January 18, 2024
Revision received October 28, 2024
Accepted October 30, 2024 ▪
EQUITABLE SHIFTS IN YOUTH RESILIENCE? 11