School Psychology

The Curative Effect of Schools: A Longitudinal Study of the Impact of School

Climate, School Identification, and Resilience on Adolescent Mental Health
Aseel Sahib, Junwen Chen, Katherine J. Reynolds, and Diana Cárdenas
Online First Publication, September 7, 2023. https://dx.doi.org/10.1037/spq0000571

CITATION
Sahib, A., Chen, J., Reynolds, K. J., & Cárdenas, D. (2023, September 7). The Curative Effect of Schools: A Longitudinal
Study of the Impact of School Climate, School Identification, and Resilience on Adolescent Mental Health. School
Psychology. Advance online publication. https://dx.doi.org/10.1037/spq0000571
 School Psychology
© 2023 American Psychological Association
ISSN: 2578-4218 https://doi.org/10.1037/spq0000571

The Curative Effect of Schools: A Longitudinal Study of the Impact of School

Climate, School Identification, and Resilience on Adolescent Mental Health
Aseel Sahib1, Junwen Chen1, Katherine J. Reynolds1, 2, and Diana Cárdenas3
1
Department of Psychology, School of Medicine and Psychology, The Australian National University
2
Department of Education, Melbourne Graduate School of Education, The University of Melbourne
3
Department of Psychology, University of Montreal

Mental illness in adolescents is on the rise, thus it is vital to study factors that can improve youth mental health.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

The extant theory and research have identified both social (school climate; school identification) and individual
This document is copyrighted by the American Psychological Association or one of its allied publishers.

(resilience) constructs as protectors of mental health. However, these protective factors remain in silo and require
further integration. To address this issue, the present study proposed and investigated an integrative model in
which social factors (i.e., school climate, school identification) nurture individual (i.e., resilience) protective
factors, which in turn impact adolescent mental health. Using three-wave longitudinal data (2017–2019) from
school students (Grades 7–8; N = 1,357), we found evidence supporting the integrated model examining five
dimensions of mental health: anxiety, depression, happiness, life satisfaction, and positive affect. Greater Wave 1
school climate predicted greater identification 1 year later, which in turn predicted greater resilience.
Furthermore, greater resilience predicted lower depression and anxiety, and greater happiness, life satisfaction,
and positive affect 1 year later. These results support efforts to strengthen the school climate and reconceptualize
resilience as an outcome of social processes.

Impact and Implications

School-based resilience programs can improve youth mental health through school climate and school
identification. Our integrated model demonstrates that a group environment (school climate) helps youth
to build psychological (school) identification and resilience, which, in turn, promotes greater mental
health. Resilience has often been conceptualized as an individual attribute. However, our results suggest
that it can be viewed as being related to groups, which could make people feel encouraged and supported.

Keywords: adolescent mental health, school climate, school identification, resilience

Supplemental materials: https://doi.org/10.1037/spq0000571.supp

Mental illnesses such as anxiety and depression have an increased (the presence or absence of a disease involving psychological
incidence and prevalence rate during adolescence (Kessler et al., 2007) symptoms such as anxiety and depression) and well-being (the extent
and have been linked to increases in risk-taking behavior (i.e., smoking to which an individual feels confident and eager about life; typically
and drinking) and self-injurious behavior (Herpertz-Dahlmann et al., comprised of life satisfaction, positive affect, and [lack of ] negative
2013). To develop better mental health strategies and programs, it is affect; Diener, 2000). One such factor is resilience. Resilient youth—
crucial to identify factors impacting youth mental health. Specifically, that is, those who can overcome adversity—show better mental health.
it is vital that we investigate factors that influence mental illness Resilience is usually conceptualized as an individual attribute, where
the locus of explanation is young people themselves and their specific
characteristics. However, these individual characteristics could be
nourished by social factors. This study goes beyond an individual focus
and investigates an integrated model of resilience in which school
Aseel Sahib https://orcid.org/0000-0001-6403-3977
social factors are critical for resilience. Thus, we conceptualize
This work was supported by the Australian Capital Territory (ACT)
Education Directorate (Grant QPY200-31) to Katherine J. Reynolds. The
resilience as the outcome of school social psychological processes and
authors would like to acknowledge that secondary data were used in this study. propose that the schools’ social environment (school climate) builds
The data were collected by the ACT Education Directorate in partnership with psychological school identification (or connectedness) and in turn
Australian National University (School of Medicine and Psychology) resilience, which, then, promotes greater mental health (i.e., lower
researchers. The grand project was called The Australian School Climate and anxiety and depression symptoms and greater well-being).
School Identification Measurement Tool. The project has been running yearly
since 2007. The authors also would like to thank the team at the ACT Education
Directorate for allowing the authors to use the data collected. The authors have no Resilience
conflicts of interest to disclose.
Correspondence concerning this article should be addressed to Aseel Originally conceptualized as a personality trait contributing to
Sahib, Department of Psychology, School of Medicine and Psychology, children’s mental health, resilience is now conceptualized as the
The Australian National University, Building 39, Science Road, Canberra, interaction between a person and their environment, where internal
ACT 2601, Australia. Email: aseel.sahib@anu.edu.au (e.g., personality and worldviews) and external resources (e.g., the

1
 2 SAHIB, CHEN, REYNOLDS, AND CÁRDENAS

social environment) are utilized to overcome an aversive condition and through the process of social identification that groups, and their
achieve positive outcomes (Fergus & Zimmerman, 2005). Moreover, resources, become internalized (Jetten et al., 2017). In the school
there is a growing awareness that resilience is a process, that is, that context, school identification reflects the internalization of the school
individuals need to recognize, access, and act on the support provided as a relevant group of reference (e.g., students think of themselves as
to maintain mental health. When such support is unavailable or members of the school, Bizumic et al., 2009; Reynolds et al., 2017)
underutilized, well-being decreases while a mental illness is likely to and potentially its resources for resilience. In a two-wave study, school
develop (e.g., becoming depressed after losing a job). In contrast, identification was shown to be negatively associated with mental
when resources are available and used, this can protect against health (depressive and anxious symptoms) in adolescents (Shochet
adversity, preventing mental illness while achieving well-being. et al., 2010). Specifically, when prior symptoms were controlled for,
To date, research has mostly focused on individual-level factors lower school identification predicted increased depressive symptoms
as the resources of resilience (e.g., self-esteem; Arslan, 2019). While 1 year later and anxiety symptoms (for girls).
we do not dispute the importance of these factors, we posit that Importantly, a positive school climate can foster school identifica-
social factors, and particularly the groups to which individuals tion because groups with positive characteristics facilitate social
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

belong, are also important sources of resources (S. A. Haslam et al., identification (Cárdenas & de la Sablonnière, 2020). This is based on
This document is copyrighted by the American Psychological Association or one of its allied publishers.

2018) to be considered in the resilience process. For adolescents, the social identity approach (Tajfel & Turner, 1979), which postulates
schools are meaningful groups that significantly affect their health that individuals are motivated to belong to groups that are positively
and development. To investigate this, we examined schools as social valued. In schools, previous longitudinal research has indeed shown
groups and how they contribute to the resilience process, protecting that a positive school climate predicted greater social identification
and promoting mental health. 1 year later (Klik et al., 2023). Thus, we propose that school climate is
an external social factor that facilitates school identification. With
these external resources internalized, students can experience greater
School Climate, School Identification, and Mental Health
resilience and in turn better mental health (see Figure 1).
Group membership is an important facet of human life (Tajfel &
Turner, 1979; J. C. Turner et al., 1987). A group that significantly
affects adolescents’ health and development is their school. Schools A Fragmented Model
can be viewed as social groups with specific group characteristics Thus far, evidence for this model is fragmented. Some research has
that define them and combine to form a school climate (Reynolds examined school social factors but has ignored resilience. In a series of
et al., 2017). The term school climate captures the shared values and studies, school climate predicted greater school identification, which,
approach, academic emphasis, and positive relations with other in turn, predicted a range of positive school outcomes including lower
school members (Lee et al., 2017). School climate is associated with school aggression and victimization (I. Turner et al., 2018). Cross-
including lower psychopathology and greater well-being in youth sectional research in Australia (Bizumic et al., 2009) and China (Tong
(e.g., Rathmann et al., 2018; Suldo et al., 2012). For example, a et al., 2019) also found that school identification mediates the
systematic review found positive associations between school relationship between school climate and adolescent mental health.
climate and student mental health (Aldridge & McChesney, 2018). Other studies have examined school social factors and resilience
In one rare longitudinal study, teacher-reported school climate was together but have ignored the internalization process. In two cross-
associated with lower depressive symptoms in youth within the sectional studies, Riekie et al. (2017) and Aldridge et al. (2020) found
same year and over time (Pössel et al., 2016). Thus, there is evidence that resilience acted as a mediator between school climate and well-
that school climate protects and promotes mental health over time. being. Similarly, Scarf et al. (2017) employed an experimental
We argue that school climate can be considered a social resource paradigm and found that adolescents in a condition that fostered a
to be used in the resilience process. Groups that have positive
characteristics are well suited to provide individuals with a wide
array of resources (Hobfoll et al., 1995), including certainty, safety Figure 1
(Ketturat et al., 2016), and social support (Jetten et al., 2017). These The Integrated Model
resources can mitigate the effect of negative events and/or maximize
School
the positive outcomes of positive events (Jones & Jetten, 2011). In Resilience
Identification
schools, a positive school climate means that the social environment 2018
2018
is supporting positive relations between members, academic success,
and clear goals and instructions. These psychological resources are
School Mental
embedded within the school climate of a school. Thus, schools with a
Climate Health
positive school climate can be considered resourceful groups that
2017 2019
offer students care and support when encountering difficulties.
Mental
However, for school climate to be a useful resource, individuals
first need to consider the group as self-relevant and meaningful (i.e., Health

internalize the group into one’s identity; J. C. Turner et al., 1987). Sex

This process is described by the social identity approach (Tajfel & Grade

Turner, 1979) to health (Jetten et al., 2017), which postulates that 2017

when groups are psychologically internalized as part of self, that is,

when individuals identify with their groups, they are aware of and Note. The integrated model with temporal information. 2017 is Wave 1,
can access the resources offered by their groups. In other words, it is 2018 is Wave 2, and 2019 is Wave 3.
 CLIMATE, IDENTIFICATION, RESILIENCE ON MENTAL HEALTH 3

sense of belonging with a school voyage group had greater resilience 22.4%; no data of race or ethnicity were collected). School
over time. Despite these fragmented advances, an integrated resilience socioeconomic scores were measured with the Index of Community
model, one that accounts for school climate, school identification, Socio-Educational Advantage, a nationally standardized value with
resilience, and mental health, remains untested. an average of 1,000 and higher values indicating greater advantage
(Australian Curriculum Assessment & Reporting Authority, 2014).
In our sample, the average school socioeconimic status was 1082.17
The Present Study
(SD = 50.06).
The present study proposed and tested an integrated model in
which a positive school climate facilitates greater identification with
Measures
the school, which enables resilience and in turn mental health—
lower levels of anxiety and depression symptoms and higher levels McDonald’s omega for all measures in this study can be found in
of well-being (see Figure 1). We tested this integrated model using Supplemental Table S1.
a longitudinal design including three waves of data collection.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Specifically, we hypothesized that:

This document is copyrighted by the American Psychological Association or one of its allied publishers.

School Climate at Wave 1 and School Identification

at Wave 2
Hypothesis 1: Positive school climate in Wave 1 predicted
lower levels of anxiety and depression symptoms and higher To measure school climate and school identification, a shortened
levels of well-being in Wave 3. version of the Australian School Climate and School Identification
Measurement Tool (SCASIM-St; Lee et al., 2017) was employed.
Hypothesis 2: Higher school identification and higher levels This scale simultaneously measured school climate and school
of resilience in Wave 2 predicted lower levels of anxiety and identification. The school climate dimensions that were measured
depression symptoms and higher levels of well-being in include shared values and approach (e.g., “Students and staff are
Wave 3. working toward the same goals”), academic emphasis (e.g.,
Hypothesis 3: Wave 2 school identification and resilience “Teachers encourage students to try out new ideas/think indepen-
sequentially mediated the relationship between Wave 1 school dently”), student–student relations (e.g., “Students care about each
climate and Wave 3 mental health. other”), and student–staff relations (e.g., “Staff treat students with
respect”).
School identification was measured with five items such as
To test the integrated model, sex and grade (proxy of age) were
“I feel a strong connection with this school.” Respondents used a
controlled for, as previous studies demonstrate their impact on
5-point Likert scale, ranging from 1 (strongly disagree) to 5
students’ mental health (Patalay et al., 2015; Van Droogenbroeck
(strongly agree). Both school climate and school identification
et al., 2018). For example, in Van Droogenbroeck et al. (2018),
have acceptable reliability in this study (Lee et al., 2017). The
girls reported significantly higher scores than boys and older boys
validity and reliability of the SCASIM-St in an adolescent
experienced greater psychological distress than younger boys. In
population have acceptable to excellent test–retest reliability (.66
contrast, Patalay et al. (2015) found that younger adolescents had
−.75), convergent validity (.51−.95), and criterion validity (e.g.,
significantly more emotional symptoms than older peers and that
academic achievement, and depressive mood—52%−93.4%; Lee
sex did not moderate these effects. Therefore, it is important to
et al., 2017).
control the potential impact of these factors on the sample’s mental
health.
Resilience at Wave 2
Method The Brief Resilience Scale (BRS; Smith et al., 2008) was used to
assess resilience. The BRS consisted of six items on a 5-point Likert
Participants and Procedure
scale, ranging from 1 (strongly disagree) to 5 (strongly agree).
The longitudinal data were collected (from 2017 to 2019) as part Three items were positively framed (e.g., “I tend to bounce back
of an ongoing research project. Ethics was approved by the quickly after hard times”), with the remaining three items being
Australian National University Human Research Ethics Committee, reversely scored (e.g., “I have a hard time making it through stressful
Protocol 2019/574. Students in Grades 7–10 from 18 schools in an events”). Previous research has demonstrated that BRS has high
Australian school jurisdiction participate in a yearly school climate convergent (e.g., coping, social relationships) and discriminant
and satisfaction survey, where they respond to the measures each validity (e.g., physical symptoms, depression, and anxiety; Smith
year. The education department distributed the survey and students et al., 2008).
answered during class time (online) in early September. As the
survey was deemed to be low risk, parental consent was not Anxiety and Depressive Symptoms at Waves 1 and 3
considered necessary. Only students who assented to take part in the
survey participated. The anxiety measure consisted of nine items (on a 3-point Likert
This study follows the cohort of students in Grades 7 and 8 in scale ranging from 1 [not true] to 3 [definitely true]) of the Worry
2017 (N = 1,357), with a mean age of 12.96 in the first wave (SD = subscale from the Screen for Child Anxiety-Related Emotional
0.76). Half of the students were female (n = 673, 49.6%) and in Year Disorders (Birmaher et al., 1996). Screen for Child Anxiety-Related
7 (n = 681, 50.2%). Less than a quarter of students spoke a Emotional Disorders demonstrates good test–retest reliability (.7–.9)
second language at home (proxy for cultural diversity; n = 272, and discriminative validity (e.g., depression and disruptive disorders,
 4 SAHIB, CHEN, REYNOLDS, AND CÁRDENAS

and different types of anxiety; Birmaher et al., 1997) in a child and health (anxiety, depression, happiness, life satisfaction, and positive
adolescent population. An example item is “People tell me I worry affect scores), in turn, creating the integrated model (see more
too much.” information in Appendix A). Model 1 added covariates. Student
The depression measure includes 10 items (on a 5-point Likert perceptions of Wave 1 school climate were added in Model 2
scale, ranging from 1 [does not apply to me at all] to 5 [definitely (testing Hypothesis 1), while Wave 2 school identification and
applies to me]) from the revised Boston version of the Center for Wave 2 resilience were added in Model 3 (testing Hypothesis 2).
Epidemiological Studies–Depression scale (CES-D-10; Kohout Model 4 introduced the path from school climate to school
et al., 1993). Eight of the 10 items were positively scored (two were identification, and school climate to resilience (to test the pathway
recoded), with an example item being “I felt everything I did was an between the independent variable and the mediators). Model 5
effort.” The CES-D-10 has shown good test–retest reliability (.85), included the path from school identification to resilience (to test the
convergent validity (.91), and divergent validity (.89; Björgvinsson pathway between the mediators and thus the integrated model). We
et al., 2013). tested the mediation by examining the indirect effect from school
climate to mental health via school identification and, in turn,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

resilience (the integrated model testing Hypothesis 3; see Figure 1).

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Well-Being at Waves 1 and 3

This series of models was repeated for each outcome variable. In line
Three indicators of well-being were employed: happiness, life with recommendations (Hu & Bentler, 1999), the path analysis
satisfaction, and positive affect. Happiness and life satisfaction were models were judged as acceptable if the comparative fit index (CFI)
both single-item questions. For happiness, participants answered the was close to .90, and the root-mean-square error of approximation
following item: “Taking all things together, how happy are you?” on (RMSEA) was <.08 (or <.10 for marginal fit; Hu & Bentler, 1999).
a 4-point Likert scale ranging from 1 (not at all happy) to 4 (very We used maximum likelihood and weighted least squares estimators
happy). Life satisfaction was gauged through the item “All things for our models in Mplus (Muthén & Muthén, 2023).
considered, how satisfied are you with your life as a whole these
days?” on a 5-point Likert scale ranging from 1 (very dissatisfied) to Results
5 (very satisfied). Single items for happiness (Abdel-Khalek, 2006;
Moldovan, 2017) and life satisfaction (Cheung & Lucas, 2014; Preliminary Analysis
Jovanović, 2016) demonstrate good concurrent and convergent Preliminary analyses were performed to examine patterns of
validity, respectively, with other measures such as the Oxford missing data, outliers, and assumptions of normality and multi-
Happiness Inventory (r = .63, p < .001; Abdel-Khalek, 2006) and collinearity. Details can be found in Supplemental Tables S1 and S3,
the Satisfaction With Life Scale (r = .67, p < .001; Jovanović, 2016). with critical analyses discussed here. To handle missing data, full
Positive affect was measured using the Positive Affect subscale of information maximum likelihood in Mplus was employed (Hox,
the Mental Health Inventory for Adolescents (Heubeck & Neill, 1999). The data set contained 12 univariate outliers and 28
2000). The 10 items were measured with a 6-point Likert scale multivariate outliers. For the univariate outliers, there were 11 for
ranging from 1 (none of the time) to 6 (all the time). An example item school climate 2017 and one for depression 2017. The school
is “Relaxed and free of tension.” The original scale has good internal climate outliers had Z scores lower than −3.21, reflecting a very low
consistency test–retest reliability (.70 stability over a 10-week school climate. The depression outlier had a Z score greater than
period; Heubeck & Neill, 2000) and criterion validity (Marques et al., 3.21, indicating a high depressive score. We ran analysis with and
2011) in an adolescent population. without the univariate outliers windsorized, and the results were
similar; thus, we retained the original scores. Analyses with and
Control Variables without the multivariate outliers revealed no important changes and
so they were retained (Leys et al., 2019).
Students’ grade (Year 7 and 8) and sex (male and female) as well
as well-being/mental health at Wave 1 (2017) were included as
Measurement Model
covariates.
The proposed model has six latent variables (school climate,
school identification, resilience, anxiety, depression, and positive
Statistical Analysis Plan
affect; happiness and life satisfaction are single-item variables).
Mplus 8.4 was used to build structural equation models to test if the Thus, it was first examined whether these formed six separate latent
indirect effect of school climate at Wave 1 on mental health at Wave 3 constructs. In line with Lee et al. (2017), school climate was
passed through school identification and resilience at Wave 2. This conceptualized as a second-order latent variable composed of four
analysis allows us to investigate the potential effects of our predictor first-order latent variables (academic emphasis, values and approach,
variables on adolescent mental health scores by explaining between- student–staff relations, student–student relations), with the remain-
person differences (Lüdtke & Robitzsch, 2021), as opposed to other ing variables being first-order latent variables. The fit indices of this
analyses that capture within-person temporary fluctuations (e.g., a model were above acceptable standards, χ2(1,755) = 5663.936, p <
cross-lag analysis; multilevel modeling [MLM]). The predictor .001; RMSEA = .040 [.039–.041]; CFI = .924 (see Supplemental
variables (school climate, school identification, and resilience) were Table S4 for further details). Thus, this measurement model was
added step by step, to investigate their effect on students’ mental used for structural equation models analysis.
 CLIMATE, IDENTIFICATION, RESILIENCE ON MENTAL HEALTH 5

Figure 2
Integrated Model: Anxiety

Wave 2 school Wave 2

identification resilience
B = 0.148***

B = 0.549***
B = -0.203***
B = 0.067*
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

B = 0.000

Wave 1 school
Wave 3 anxiety
climate
B = -0.034

B = -0.017***

Note. The figure represents the results of the final model (Model 5) in a figure form. The straight arrows from the predictor
variables (school climate, school identification, and resilience) to the outcome variable represent the results of the direct effect
(testing Hypotheses 1 and 2). The straight arrows from school climate to school identification, school climate to resilience, and
school identification to resilience represent the pathways needed to test the indirect effect. The dash arrow represents the
indirect effect results (from school climate to mental health via first school identification and then resilience; Hypothesis 3).
B = standardized beta coefficient.
* p < .05. *** p < .001.

Main Results. was only supported for resilience). In Model 4, the pathways from
Anxiety. Figure 2 has the final model details for anxiety (see school climate to school identification and school climate to resilience
Supplemental Table S5 for further details). In Model 1, Wave 1 were both significant. Likewise, the pathway from school identification
anxiety and gender scores significantly predicted Wave 3 anxiety to resilience (Model 5) was significant. This last model fits the data
scores. In Model 2, Wave 1 school climate did not significantly well. The indirect effect of school climate to depression via school
predict Wave 3 anxiety (Hypothesis 1 was not supported). In Model 3, identification and in turn resilience was significant. Greater Wave 1
Wave 2 school identification and resilience were added, and only school climate predicted greater identification 1 year later, which in
resilience predicted anxiety (Hypothesis 2 was only supported for turn predicted greater resilience and resulted in lower depression 1 year
resilience). In Model 4, Wave 1 school climate predicted greater later, supporting the hypothesized resilience process (Hypothesis 3 was
Wave 2 school identification and resilience. In Model 5, the path from supported). The percentage of variance accounted for by our model for
Wave 2 school identification to Wave 2 resilience was significant. depression was 2.70%.
This last model demonstrated appropriate fit indices (see model Happiness. Figure 4 has the final model details for happiness (see
statistics in Supplemental Table S2). In line with Hypothesis 3 (the Supplemental Table S7 for further details). In Model 1, Wave 1
hypothesized resilience process), the indirect effect of school climate happiness and gender scores significantly predicted Wave 3 happiness
to anxiety via school identification and in turn resilience was significant. scores. In Model 2, Wave 1 school climate significantly predicted
Greater Wave 1 school climate predicted greater identification 1 year Wave 3 happiness (Hypothesis 1 was supported). Similarly, when
later, which in turn predicted greater resilience and resulted in lower Wave 2 school identification and resilience (Model 3) were added,
anxiety 1 year later (see Appendix B for more information). The happiness was significantly predicted by both (Hypothesis 2 was
percentage of variance accounted for by our model for anxiety was supported). In Model 4, the pathways from school climate to school
7.80%. identification and school climate to resilience were both significant.
Depression. Figure 3 has the final model details for depression Likewise, the path from school identification to resilience (Model 5)
(see Supplemental Table S6 for further details). In Model 1, Wave 1 was significant. This last model fits the data well. Testing mediation
depression and gender scores significantly predicted Wave 3 effects showed that the indirect effect of school climate to happiness
depression scores. Depression was significantly predicted by school via school identification and in turn resilience was significant. Greater
climate in Model 2 (Hypothesis 1 was supported). In contrast, when Wave 1 school climate predicted greater identification 1 year later,
Wave 2 school identification and resilience (Model 3) were added, which in turn predicted greater resilience. In turn, greater resilience
depression was significantly predicted by resilience only (Hypothesis 2 predicted greater happiness 1 year later, offering support for the
 6 SAHIB, CHEN, REYNOLDS, AND CÁRDENAS

Figure 3
Integrated Model: Depression

Wave 2 school Wave 2

identification resilience
B = 0.140***

B = 0.553***
B = -0.019***
B = 0.065
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

B = 0.014

Wave 1 school Wave 3

climate depression
B = 0.036

B = -0.019***

Note. The figure represents the results of the final model (Model 5) in a figure form. The straight arrows from the predictor
variables (school climate, school identification, and resilience) to the outcome variable represent the results of the direct effect
(testing Hypotheses 1 and 2). The straight arrows from school climate to school identification, school climate to resilience, and
school identification to resilience represent the pathways needed to test the indirect effect. The dash arrow represents the
indirect effect results (from school climate to mental health via first school identification and then resilience; Hypothesis 3).
B = standardized beta coefficient.
*** p < .001.

Figure 4
Integrated Model: Happiness

Wave 2 school Wave 2

identification resilience
B = 0.142***

B = 0.555***
B = 0.150***
B = 0.308***

B = 0.013

Wave 1 school
Wave 3 happiness
climate
B = 0.040

B = 0.012***

Note. The figure represents the results of the final model (Model 5) in a figure form. The straight arrows from the predictor
variables (school climate, school identification, and resilience) to the outcome variable represent the results of the direct effect
(testing Hypotheses 1 and 2). The straight arrows from school climate to school identification, school climate to resilience, and
school identification to resilience represent the pathways needed to test the indirect effect. The dash arrow represents the
indirect effect results (from school climate to mental health via first school identification and then resilience; Hypothesis 3).
B = standardized beta coefficient.
*** p < .001.
 CLIMATE, IDENTIFICATION, RESILIENCE ON MENTAL HEALTH 7

hypothesized resilience process (Hypothesis 3 was supported). The positive affect (Hypothesis 1 was supported). Similarly, when Wave
percentage of variance accounted for by our model for happiness was 2 school identification and resilience (Model 3) were added, positive
16.10%. affect was significantly predicted by both (Hypothesis 2 was
Life Satisfaction. Figure 5 has the final model details for life supported). In Model 4, both the pathways from school climate
satisfaction (see Supplemental Table S8 for further details). In to school identification and school climate to resilience were
Model 1, Wave 1 life satisfaction and gender scores significantly significant, as was the pathway from school identification to
predicted Wave 3 life satisfaction scores. In Model 2, Wave 1 resilience in Model 5. This last model fits the data well. In addition,
school climate significantly predicted Wave 3 life satisfaction the indirect effect of school climate to positive affect via school
(Hypothesis 1 was supported). Similarly, when Wave 2 school identification and in turn resilience was significant. Greater Wave 1
identification and resilience (Model 3) were added, life satisfaction school climate predicted greater identification 1 year later, which in
was significantly predicted by both (Hypothesis 2 was supported). turn predicted greater resilience. Greater resilience predicted greater
In Model 4, both the pathways from school climate to school positive affect 1 year later, supporting the hypothesized resilience
identification and school climate to resilience were significant. process (Hypothesis 3 was supported). The percentage of variance
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Likewise, the pathway from school identification to resilience accounted for by our model for positive affect was 23.40%.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

(Model 5, which fits the data well) was significant. Moreover, the
indirect effect of school climate to life satisfaction via school
Discussion
identification and in turn resilience was significant. Greater Wave 1
school climate predicted greater identification 1 year later, which in This study aimed to test an integrated model of resilience by
turn predicted greater resilience and resulted in greater life emphasizing the impact of school climate and school identification on
satisfaction 1 year later. This provides support to the expected the resilience process. The tested models found a direct effect of
resilience process (Hypothesis 3 was supported). The percentage of school climate and of school identification on well-being but not
variance accounted for by our model for life satisfaction was anxiety and depression, partially confirming Hypotheses 1 and 2. This
24.10%. may be because elements of school climate and identification increase
Positive Affect. Figure 6 has the final model details for positive the way activities are interpreted and experienced within schools,
affect (see Supplemental Table S9 for further details). In Model 1, all which promotes positive well-being (Lombardi et al., 2019; Van Dick
covariates significantly predicted Wave 3 positive affect scores. In et al., 2017), but not necessarily protects against psychopathology.
Model 2, Wave 1 school climate significantly predicted Wave 3 Multiple risk factors for adolescent anxiety and depression are outside

Figure 5
Integrated Model: Life Satisfaction

Wave 2 school Wave 2

identification resilience
B = 0.141***

B = 0.555***
B = 0.160***
B = 0.252***

B = 0.017

Wave 1 school
Wave 3 life satisfaction
climate
B = 0.060

B = 0.013***

Note. The figure represents the results of the final model (Model 5) in a figure form. The straight arrows from the predictor
variables (school climate, school identification, and resilience) to the outcome variable represent the results of the direct effect
(testing Hypotheses 1 and 2). The straight arrows from school climate to school identification, school climate to resilience, and
school identification to resilience represent the pathways needed to test the indirect effect. The dash arrow represents the
indirect effect results (from school climate to mental health via first school identification and then resilience; Hypothesis 3).
B = standardized beta coefficient.
*** p < .001.
 8 SAHIB, CHEN, REYNOLDS, AND CÁRDENAS

the school’s control, including temperament, childhood trauma, and for youth mental health. While there are interventions that use social
parent–child interactions (Thapar et al., 2012; Valance & Fernandez, identities to reduce mental illness in adults (e.g., Groups 4 Health;
2016). The school as a group may be unable to fully account for C. Haslam et al., 2019), more work is required to validate school-
mental ill-health because doing so requires integrating individual and based resilience programs that strengthen existing climate and
family factors (Werner-Seidler et al., 2017). identification and to promote greater resilience and mental health.
However, as expected, ill-health and well-being indicators were
significantly predicted by school climate through first greater school
identification and then greater resilience (Hypothesis 3). Thus, Limitations, Constraints on Generality, and
school climate may be a valuable resource to be used in the process Future Directions
of resilience because these resources are internalized via school The longitudinal design, consistency of findings across several
identification such that students experience a connection to their mental health indicators, large sample size, and the inclusion of
school. In turn, the internalization of the group (along with access important covariates are among the strengths of this research. Despite
and use of resources) fosters resilience and leads to improvement
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

this, there are important caveats. While the sample size was large, it
in mental health. Thus, this study extends our understanding of
This document is copyrighted by the American Psychological Association or one of its allied publishers.

was not possible to run many control variables in one analysis.

resilience beyond individual factors to formally integrate the role of Likewise, we did not have enough schools to run multilevel analyses,
social school factors. which would have granted us the ability to cluster standard errors at
Interestingly, resilience alone did not significantly mediate the the school level. Moreover, the region in which the study was
relationship between school climate and any mental health indicator. conducted is homogeneous in terms of socioeconomic status and
This suggests that schools need to be psychologically meaningful education, although it is like the ethnic diversity of Australia. Future
(i.e., school identification) for school climate to be a source of research should include a more heterogeneous sample involving
resilience. The integration of external resources (i.e., school climate) different ethnicities and migration backgrounds to ensure the
via social identification for resilience is a clear contribution to generalizability of the results. Additionally, multiple-item questions
education and clinical psychology. Our additional analyses provide to measure happiness and life satisfaction would have been better
further evidence that greater school climate and school identification suited to assess these complex constructs. Likewise, the BRS does
strengthen resilience in time (see Supplemental Table S10). not measure adversity, a key aspect of resilience (Fergus &
Interventions that promote a positive school climate and school Zimmerman, 2005). Future research should include the experience of
identification may hence have important downstream implications life stressors. Last, the amount of variance accounted for by each

Figure 6
Integrated Model: Positive Affect

Wave 2 school Wave 2

identification resilience
B = 0.139***

B = 0.553***
B = 0.149***
B = 0.247***

B = 0.018

Wave 1 school Wave 3

climate positive affect
B = 0.049

B = 0.012***

Note. The figure represents the results of the final model (Model 5) in a figure form. The straight arrows from the predictor
variables (school climate, school identification, and resilience) to the outcome variable represent the results of the direct effect
(testing Hypotheses 1 and 2). The straight arrows from school climate to school identification, school climate to resilience, and
school identification to resilience represent the pathways needed to test the indirect effect. The dash arrow represents the
indirect effect results (from school climate to mental health via first school identification and then resilience; Hypothesis 3).
B = standardized beta coefficient.
*** p < .001.
 CLIMATE, IDENTIFICATION, RESILIENCE ON MENTAL HEALTH 9

model was low, although this is consistent with other school climate Cicchetti, D. (1994). Guidelines, criteria, and rules of thumb for evaluating
studies (e.g., 13% in Modin et al., 2011). Future studies should normed and standardised assessment instruments in psychology.
consider other variables that could contribute further to the findings Psychological Assessment, 6(4), 284–290. https://doi.org/10.1037/1040-
(e.g., developmental–behavioral challenges and family factors like 3590.6.4.284
parental support and parent–child conflict; Kieling et al., 2011; King D’haenens, E., Van Damme, J., & Onghena, P. (2010). Multilevel
exploratory factor analysis: Illustrating its surplus value in educational
& Merchant, 2008).
effectiveness research. School Effectiveness and School Improvement,
21(2), 209–235. https://doi.org/10.1080/09243450903581218
Conclusion Diener, E. (2000). SWB: The science of happiness, and a proposal for a
national index. American Psychologist, 55(1), 34–43. https://doi.org/10
The aim of this study was to investigate an integrated framework of .1037/0003-066X.55.1.34
resilience that considers group resources (school climate), and their Fergus, S., & Zimmerman, M. A. (2005). Adolescent resilience: A
internalization (school identification), to foster resilience and, in turn, framework for understanding healthy development in the face of risk.
mental health in the adolescent population. In doing so, it addresses a Annual Review of Public Health, 26(1), 399–419. https://doi.org/10.1146/
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

crucial gap in clinical and educational research, providing pathways annurev.publhealth.26.021304.144357

This document is copyrighted by the American Psychological Association or one of its allied publishers.

to interventions for school-based resilience programs. Haslam, C., Cruwys, T., Chang, M. X., Bentley, S. V., Haslam, S. A., Dingle,
G. A., & Jetten, J. (2019). GROUPS 4 HEALTH reduces loneliness and
social anxiety in adults with psychological distress: Findings from a
References randomized controlled trial. Journal of Consulting and Clinical
Psychology, 87(9), 787–801. https://doi.org/10.1037/ccp0000427
Abdel-Khalek, A. (2006). Measuring happiness with a single-item scale. Haslam, S. A., McMahon, C., Cruwys, T., Haslam, C., Jetten, J., & Steffens,
Social Behavior and Personality, 34(2), 139–150. https://doi.org/10.2224/ N. K. (2018). Social cure, what social cure? The propensity to underestimate
sbp.2006.34.2.139
the importance of social factors for health. Social Science & Medicine, 198,
Aldridge, J. M., & McChesney, K. (2018). The relationships between school
14–21. https://doi.org/10.1016/j.socscimed.2017.12.020
climate and adolescent mental health and wellbeing: A systematic literature
Herpertz-Dahlmann, B., Bühren, K., & Remschmidt, H. (2013). Growing up is
review. International Journal of Educational Research, 88, 121–145.
hard: Mental disorders in adolescence. Deutsches Ärzteblatt International,
https://doi.org/10.1016/j.ijer.2018.01.012
110(25), 432–439. https://doi.org/10.3238/arztebl.2013.0432
Aldridge, J. M., McChesney, K., & Afari, E. (2020). Associations between
Heubeck, B. G., & Neill, J. T. (2000). Confirmatory factor analysis and reliability
school climate and student life satisfaction: Resilience and bullying as
of the Mental Health Inventory for Australian adolescents. Psychological
mediating factors. Learning Environments Research, 23(1), 129–150.
Reports, 87(2), 431–440. https://doi.org/10.2466/pr0.2000.87.2.431
https://doi.org/10.1007/s10984-019-09296-9
Hobfoll, S. E., Briggs, S., & Wells, J. (1995). Community stress and
Arslan, G. (2019). Mediating role of the self–esteem and resilience in the
resources: Actions and reactions. In S. E. Hobfoll & M. W. deVries (Eds.),
association between social exclusion and life satisfaction among adolescents.
Extreme stress and communities: Impact and intervention (pp. 137–158).
Personality and Individual Differences, 151, Article 109514. https://doi.org/
Kluwer Academic/Plenum Publishers. https://doi.org/10.1007/978-94-
10.1016/j.paid.2019.109514
015-8486-9_6
Australian Curriculum Assessment and Reporting Authority. (2014). What
Hox, J. (1999). A review of current software for handling missing data.
does the ICSEA value mean? https://acaraweb.blob.core.windows.net/re
Kwantitatieve Methoden, 62, 123–138. https://www.vvsor.nl/wp-content/
sources/About_icsea_2014.pdf
uploads/2020/06/KM1999062009.pdf
Birmaher, B., Khetarpal, S., Brent, D., Cully, M., Balach, L., Kaufman, J., &
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance
Neer, S. M. (1997). The screen for child anxiety related emotional
structure analysis: Conventional criteria versus new alternatives.
disorders (SCARED): Scale construction and psychometric characteristics.
Structural Equation Modeling, 6(1), 1–55. https://doi.org/10.1080/
Journal of the American Academy of Child & Adolescent Psychiatry,
36(4), 545–553. https://doi.org/10.1097/00004583-199704000-00018 10705519909540118
Birmaher, B., Ryan, N. D., Williamson, D. E., Brent, D. A., Kaufman, J., Jetten, J., Haslam, S. A., Cruwys, T., Greenaway, K. H., Haslam, C., &
Dahl, R. E., Perel, J., & Nelson, B. (1996). Childhood and adolescent Steffens, N. K. (2017). Advancing the social identity approach to health
depression: A review of the past 10 years. Part I. Journal of the American and well-being: Progressing the social cure research agenda. European
Academy of Child & Adolescent Psychiatry, 35(11), 1427–1439. https:// Journal of Social Psychology, 47(7), 789–802. https://doi.org/10.1002/
doi.org/10.1097/00004583-199611000-00011 ejsp.2333
Bizumic, B., Reynolds, K. J., Turner, J. C., Bromhead, D., & Subasic, E. Jones, J. M., & Jetten, J. (2011). Recovering from strain and enduring pain:
(2009). The role of the group in individual functioning: School identification Multiple group memberships promote resilience in the face of physical
and the psychological well-being of staff and students. Applied Psychology, challenges. Social Psychological & Personality Science, 2(3), 239–244.
58(1), 171–192. https://doi.org/10.1111/j.1464-0597.2008.00387.x https://doi.org/10.1177/1948550610386806
Björgvinsson, T., Kertz, S. J., Bigda-Peyton, J. S., McCoy, K. L., & Aderka, Jovanović, V. (2016). The validity of the Satisfaction with Life Scale in
I. M. (2013). Psychometric properties of the CES-D-10 in a psychiatric adolescents and a comparison with single-item life satisfaction measures:
sample. Assessment, 20(4), 429–436. https://doi.org/10.1177/107319111348 A preliminary study. Quality of Life Research, 25(12), 3173–3180. https://
1998 doi.org/10.1007/s11136-016-1331-5
Cárdenas, D., & de la Sablonnière, R. (2020). Participating in a new group Kessler, R. C., Angermeyer, M., Anthony, J. C., De Graaf, R., Demyttenaere,
and the identification processes: The quest for a positive social identity. K., Gasquet, I., De Girolamo, G., Gluzman, S., Gureje, O., Haro, J. M.,
British Journal of Social Psychology, 59(1), 189–208. https://doi.org/10 Kawakami, N., Karam, A., Levinson, D., Medina Mora, M. E., Oakley
.1111/bjso.12340 Browne, M. A., Posada-Villa, J., Stein, D. J., Adley Tsang, C. H., Aguilar-
Cheung, F., & Lucas, R. E. (2014). Assessing the validity of single-item life Gaxiola, S., … Ustün, T. B. (2007). Lifetime prevalence and age-of-onset
satisfaction measures: Results from three large samples. Quality of Life distributions of mental disorders in the World Health Organization’s world
Research, 23(10), 2809–2818. https://doi.org/10.1007/s11136-014-0726-4 mental health survey initiative. World Psychiatry, 6(3), 168–176.
 10 SAHIB, CHEN, REYNOLDS, AND CÁRDENAS

Ketturat, C., Frisch, J. U., Ullrich, J., Häusser, J. A., van Dick, R., & study. School Mental Health, 8(4), 425–440. https://doi.org/10.1007/
Mojzisch, A. (2016). Disaggregating within- and between-person effects s12310-016-9191-2
of social identification on subjective and endocrinological stress reactions Rathmann, K., Herke, M., Heilmann, K., Kinnunen, J. M., Rimpelä, A.,
in a real-life stress situation. Personality and Social Psychology Bulletin, Hurrelmann, K., & Richter, M. (2018). Perceived school climate,
42(2), 147–160. https://doi.org/10.1177/0146167215616804 academic well-being and school-aged children’s self-rated health: A
Kieling, C., Baker-Henningham, H., Belfer, M., Conti, G., Ertem, I., mediator analysis. European Journal of Public Health, 28(6), 1012–1018.
Omigbodun, O., Rohde, L. A., Srinath, S., Ulkuer, N., & Rahman, A. https://doi.org/10.1093/eurpub/cky089
(2011). Child and adolescent mental health worldwide: Evidence for Reynolds, K. J., Subasic, E., Bromhead, D., & Lee, E. (2017). The school as a
action. Lancet, 378(9801), 1515–1525. https://doi.org/10.1016/S0140- group system: School climate, school identity and school outcomes. In K.
6736(11)60827-1 Mavor, M. J. Platow, & B. Bizumic (Eds.), The self, social identity and
King, C. A., & Merchant, C. R. (2008). Social and interpersonal factors relating education (pp. 55–69). Psychology Press.
to adolescent suicidality: A review of the literature. Archives of Suicide Riekie, H., Aldridge, J. M., & Afari, E. (2017). The role of the school climate
Research, 12(3), 181–196. https://doi.org/10.1080/13811110802101203 in high school students’ mental health and identity formation: A south
Klik, K. A., Cárdenas, D., & Reynolds, K. J. (2023). School climate, school Australian study. British Educational Research Journal, 43(1), 95–123.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

identification and student outcomes: A longitudinal investigation of https://doi.org/10.1002/berj.3254

This document is copyrighted by the American Psychological Association or one of its allied publishers.

student well-being. The British Journal of Educational Psychology. Salkind, N. J. (2010). Multilevel modelling. Encyclopedia of Research
Advance online publication. https://doi.org/10.1111/bjep.12597 Design, 1(10), 1473–1477. https://doi.org/10.4135/9781412961288
Kohout, F. J., Berkman, L. F., Evans, D. A., & Cornoni-Huntley, J. (1993). Two Scarf, D., Hayhurst, J. G., Riordan, B. C., Boyes, M., Ruffman, T., & Hunter,
shorter forms of the CES-D depression symptoms index. Journal of Aging J. A. (2017). Increasing resilience in adolescents: The importance of social
and Health, 5(2), 179–193. https://doi.org/10.1177/089826439300500202 connectedness in adventure education programmes. Australasian
Lee, E., Reynolds, K. J., Subasic, E., Bromhead, D., Lin, H., Marinov, V., & Psychiatry, 25(2), 154–156. https://doi.org/10.1177/1039856216671668
Smithson, M. (2017). Development of a dual school climate and school Shochet, I. M., Dadds, M. R., Ham, D., & Montague, R. (2010). School
identification measure–student (SCASIM-st). Contemporary Educational connectedness is an underemphasized parameter in adolescent mental
Psychology, 49, 91–106. https://doi.org/10.1016/j.cedpsych.2017.01.003 health: Results of a community prediction study. Journal of Clinical Child
Leys, C., Delacre, M., Mora, Y. L., Lakens, D., & Ley, C. (2019). How to and Adolescent Psychology, 35(2), 170–179. https://doi.org/10.1207/
classify, detect, and manage univariate and multivariate outliers, with s15374424jccp3502_1
emphasis on pre-registration. Revue Internationale de Psychologie Smith, B. W., Dalen, J., Wiggins, K., Tooley, E., Christopher, P., & Bernard,
Sociale, 32(1), Article 5. https://doi.org/10.5334/irsp.289 J. (2008). The brief resilience scale: Assessing the ability to bounce back.
Lombardi, E., Traficante, D., Bettoni, R., Offredi, I., Giorgetti, M., & Vernice, International Journal of Behavioral Medicine, 15(3), 194–200. https://
M. (2019). The impact of school climate on well-being experience and school doi.org/10.1080/10705500802222972
engagement: A study with high-school students. Frontiers in Psychology, 10, Suldo, S. M., McMahan, M. M., Chappel, A. M., & Loker, T. (2012).
Article 2482. https://doi.org/10.3389/fpsyg.2019.02482 Relationships between perceived school climate and adolescent mental
Lüdtke, O., & Robitzsch, A. (2021). A critique of the random intercept cross- health across genders. School Mental Health, 4(2), 69–80. https://doi.org/
lagged panel model. PsyArXiv. https://doi.org/10.31234/osf.io/6f85c 10.1007/s12310-012-9073-1
Maas, C. J., & Hox, J. J. (2005). Sufficient sample sizes for multilevel Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup
modelling. Methodology: European Journal of Research Methods for the conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of
Behavioral and Social Sciences, 1(3), 86–92. https://doi.org/10.1027/ intergroup relations (pp. 33–47). Brooks/Cole.
1614-2241.1.3.86 Thapar, A., Collishaw, S., Pine, D. S., & Thapar, A. K. (2012). Depression in
Marques, S. C., Pais-Ribeiro, J. L., & Lopez, S. J. (2011). Use of the “Mental adolescence. Lancet, 379(9820), 1056–1067. https://doi.org/10.1016/
Health Inventory—5” with portuguese 10–15 years old. The Spanish S0140-6736(11)60871-4
Journal of Psychology, 14(1), 478–485. https://doi.org/10.5209/rev_SJOP Tong, L., Reynolds, K., Lee, E., & Liu, Y. (2019). School relational climate,
.2011.v14.n1.43 social identity, and student well-being: New evidence from China on student
Modin, B., Östberg, V., Toivanen, S., & Sundell, K. (2011). Psychosocial depression and stress levels. School Mental Health, 11(3), 509–521. https://
working conditions, school sense of coherence and subjective health doi.org/10.1007/s12310-018-9293-0
complaints. A multilevel analysis of ninth grade pupils in the Stockholm Turner, I., Reynolds, K. J., Lee, E., Subasic, E., & Bromhead, D. (2018).
area. Journal of Adolescence, 34(1), 129–139. https://doi.org/10.1016/j Understanding aggression and victimization: Negative binomial model-
.adolescence.2010.01.004 ling with supportive school climate, mental health, and social identity
Moldovan, C. P. (2017). AM happy scale: Reliability and validity of a single- mediation. Translational Issues in Psychological Science, 4(4), 380–402.
item measure of happiness (Publication No. 438) [Doctoral dissertation, https://doi.org/10.1037/tps0000174
Loma Linda University of California]. TheScholarsRepository@LLU. Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S.
Muthén, L., & Muthén, B. (2023). Estimators and algorithms. https://www (1987). Rediscovering the social group: A self-categorization theory. Basil
.statmodel.com/features6.shtml Blackwell.
Patalay, P., Belsky, J., Fonagy, P., Vostanis, P., Humphrey, N., Deighton, J., Valance, A. K., & Fernandez, V. (2016). Anxiety disorders in children and
& Wolpert, M. (2015). The extent and specificity of relative age effects on adolescents: Aetiology, diagnosis, and treatment. BJPsych Advances,
mental health and functioning in early adolescence. The Journal of 22(5), 335–344. https://doi.org/10.1192/apt.bp.114.014183
Adolescent Health, 57(5), 475–481. https://doi.org/10.1016/j.jadohealth Van Dick, R., Ketturat, C., Häusser, J. A., & Mojzisch, A. (2017). Two sides
.2015.07.012 of the same coin and two routes for improvement: Integrating resilience
Pössel, P., Rakes, C., Rudasill, K. M., Sawyer, M. G., Spence, S. H., & and the social identity approach to well-being and ill-health. Health
Sheffield, J. (2016). Associations between teacher-reported school climate Psychology Open, 4(2), Article 2055102917719564. https://doi.org/10
and depressive symptoms in Australian adolescents: A 5-year longitudinal .1177/2055102917719564
 CLIMATE, IDENTIFICATION, RESILIENCE ON MENTAL HEALTH 11

Van Droogenbroeck, F., Spruyt, B., & Keppens, G. (2018). Gender differences Werner-Seidler, A., Perry, Y., Calear, A. L., Newby, J. M., & Christensen, H.
in mental health problems among adolescents and the role of social support: (2017). School-based depression and anxiety prevention programs
Results from the Belgian health interview surveys 2008 and 2013. BMC for young people: A systematic review and meta-analysis. Clinical
Psychiatry, 18(1), Article 6. https://doi.org/10.1186/s12888-018-1591-4 Psychology Review, 51, 30–47. https://doi.org/10.1016/j.cpr.2016.10.005

Appendix A
Additional Analysis

Since the data were collected from students within schools, that is, MLM is not necessary. Only positive affect had a design effect greater
that students are nested in schools, multilevel procedures are usually than 2. The model for positive affect was run with and without MLM,
used to handle their intercorrelations (D’haenens et al., 2010; Salkind, with no important differences between models (see Supplemental
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

2010). To determine whether multilevel modeling (MLM) procedures Table S2). Overall, these results suggest that there is very little variance
This document is copyrighted by the American Psychological Association or one of its allied publishers.

needed to be used for the present study, the within- and between-school to account for at the school (between) level, which may reflect a
variance of students’ 2019 outcomes were examined. The between-level homogeneous jurisdiction and the small number of schools represented
variance was not significantly different from zero (ranging from p = in the sample (18, with 20 being the recommended minimum number).
.044 to .318), indicating that there was very little variance to explain at Since there is little indication for the need to employ MLM, the models
the school level except for depression. For this variable, the model was without MLM are presented here (see Supplemental Tables S4–S8).
run with and without MLM, with no important differences between While happiness and life satisfaction were treated as continuous
models (see Supplemental Table S2). Additionally, the school intraclass variables in the main analysis, an additional analysis was performed
correlations for all outcomes were low (intraclass correlation < .04; to see whether the pattern of results was the same if the variables
Cicchetti, 1994), indicating a low correlation of errors within schools. were treated as ordinal ones. A logistic regression demonstrated
Furthermore, design effects (Maas & Hox, 2005) were lower than the similar patterns of results with school climate first predicting school
recommended standard of 2, indicating that the degree of bias in the identification, then resilience, which in turn predicted happiness
standard error is less than twice that of the expected sample error and that (b = .70, p < .000) and life satisfaction (b = .59, p < .000).

Appendix B
Model Robustness Analysis

To test whether having the two mediators measured in the same year resilience. Overall, these additional analyses support our model,
impacted our results, we conducted an additional analysis in which placing school climate and school identification as resources for
resilience in Wave 3 (the same year as the outcome variables) was used resilience and mental health. For details, see Supplemental Table S10.
as a mediator. We found the same pattern of results as described in the
main analyses, with Wave 1 school climate predicting Wave 2 school
identification, Wave 2 school identification predicting Wave 3
resilience, and Wave 3 resilience predicting Wave 3 mental health. Received October 27, 2022
Importantly, Wave 1 school climate predicted Wave 3 mental health, Revision received June 22, 2023
first through Wave 2 school identification and then through Wave 3 Accepted July 3, 2023 ▪