Positive Education: Teaching Well-Being and Resilience in Schools

Last Updated May 23, 2026

Positive education is an approach to schooling that integrates traditional academic learning with the scientific study of well-being, resilience, character, belonging, and human flourishing. It begins from a simple but far-reaching claim: schools do not merely transmit knowledge or prepare students for examinations. They shape identity, motivation, emotional development, social trust, agency, aspiration, and the conditions under which young people learn how to live with others and with themselves.

The field developed within positive psychology, but its implications reach far beyond classroom well-being exercises. Positive education asks what schooling is for, what kinds of development schools should support, and how academic learning is affected by belonging, emotional safety, meaning, relationships, institutional culture, and student agency. At its strongest, it is not a soft substitute for educational rigor. It is a broader theory of educational development.

Positive education emerged because researchers and educators increasingly recognized that emotional resilience, mental health, purpose, social relationships, motivation, and school climate play a major role in learning and long-term development. A student can be present in a classroom and still feel invisible. A school can produce test scores while weakening belonging. A curriculum can be academically demanding while leaving little room for meaning, agency, or psychological safety. Positive education asks how schools can support the whole ecology of learning.

What Is Positive Education?

Positive education applies the scientific insights of positive psychology to educational systems. Its goal is to support both academic competence and psychological well-being. This dual emphasis is central. The field does not argue that schools should stop teaching mathematics, science, literature, history, language, or the arts in favor of generalized happiness lessons. It argues that learning itself is shaped by emotional health, motivation, purpose, belonging, social relationships, institutional trust, and the wider quality of school life.

A student’s capacity to learn is not merely cognitive. It is affected by stress, sleep, nutrition, family life, disability access, peer relationships, teacher expectations, identity formation, trauma exposure, safety, belonging, confidence, and whether school feels like a place where effort can become growth. Positive education brings these conditions into the core of educational thought rather than treating them as peripheral support issues.

This means that positive education is broader than a set of classroom exercises. It includes curricula, teacher training, pastoral care, school climate, advisory systems, social-emotional learning, resilience education, character development, strengths-based reflection, and institutional norms. At its best, it also asks whether school systems themselves are organized in ways that support dignity, agency, curiosity, and development.

The concept can be summarized across several linked dimensions:

Positive education therefore reframes schooling as a developmental environment. Students are not only recipients of instruction. They are persons becoming capable of thought, relationship, responsibility, agency, and meaning. That broader view does not weaken academic seriousness. It gives academic seriousness a fuller human context.

Back to top ↑

Origins of Positive Education

The concept of positive education emerged in the early twenty-first century as researchers began exploring how the science of well-being could be integrated into school curricula and school culture. One of the most influential early formulations came from Martin Seligman and colleagues, who described positive education as education for both traditional skills and happiness. Their argument was not that happiness should replace academic learning, but that schools had a responsibility to cultivate the psychological resources that support flourishing.

The Geelong Grammar School collaboration in Australia became especially significant because it attempted a whole-school application of positive psychology rather than a narrow classroom exercise. Instead of treating well-being as a single lesson or brief intervention, the project sought to integrate positive psychology into staff training, school culture, student learning, pastoral care, and institutional practice. That whole-school ambition became a defining feature of the field.

Positive education also built on earlier intervention efforts such as the Penn Resilience Program. This program focused on cognitive and emotional skills associated with resilience, optimism, and adaptive responses to adversity. It drew partly on cognitive-behavioral principles, teaching students to identify explanatory patterns, challenge catastrophic interpretations, and respond more constructively to setbacks. Such programs helped demonstrate that some psychological skills could be taught in educational settings and that school-based interventions might influence more than short-term mood.

Over time, positive education became one of the most visible applied branches of positive psychology. Its influence grew because it addressed a real institutional need. Schools were already confronting anxiety, depression, social isolation, academic pressure, disengagement, bullying, identity stress, family instability, and widening inequality. Positive education offered a language for connecting well-being and learning rather than treating them as separate administrative domains.

Its intellectual significance lies in the fact that it connects individual psychological theory to institutional design. Once well-being enters school life, it is no longer just a personal matter. It becomes part of the educational environment itself. A student’s resilience depends partly on personal skills, but also on whether the school is safe, fair, supportive, inclusive, and developmentally responsive. Positive education becomes credible only when it attends to both sides.

Back to top ↑

PERMA and Educational Development

Many positive education programs are structured around the PERMA model of well-being, which identifies five major dimensions of flourishing: positive emotion, engagement, relationships, meaning, and accomplishment. In educational settings, these dimensions translate into concrete developmental questions. Are students emotionally supported? Are they deeply engaged in learning? Do they feel connected to peers and teachers? Can they relate their studies to purposes larger than immediate performance? Are they able to experience mastery and valued achievement?

The model gives educators a structured way to think about human development within schooling. Positive emotion matters because emotion shapes attention, receptivity, motivation, and the broader tone of learning. Engagement matters because learning depends not only on compliance but on deep involvement. Relationships matter because schools are social institutions, and belonging strongly affects development. Meaning matters because students are more likely to invest in work that feels connected to something larger than routine performance. Accomplishment matters because growth often requires visible mastery, persistence, and progress.

PERMA can also help schools avoid a one-dimensional view of student success. A student may achieve academically while feeling isolated. Another may be socially connected but disengaged from learning. Another may work hard but lack meaning or self-confidence. Another may experience positive emotion but weak mastery. A multidimensional framework allows educators to see these differences rather than collapsing student development into grades alone.

Used well, PERMA broadens what counts as educational development. It does not replace academics. It situates academic learning inside a fuller account of flourishing. Used poorly, however, PERMA can become a checklist detached from school culture. The framework is most useful when it supports deeper reflection on what students experience each day and how the institution itself shapes the conditions of learning.

Back to top ↑

How Positive Education Is Implemented

Positive education programs are often implemented by integrating well-being practices into existing curricula, advisory structures, school routines, and the broader institutional culture. These initiatives may include gratitude and reflection exercises such as the Three Good Things exercise, character strengths identification, strengths-based learning, mindfulness and emotional regulation practices, resilience training, purpose reflection, peer-support structures, restorative practices, mentoring, and social-emotional learning programs.

In many schools, these practices are introduced through classroom discussion, teacher development, pastoral systems, school assemblies, advisory periods, health curricula, or school-wide initiatives designed to support a more coherent learning environment. The best implementations tend not to treat well-being as an occasional add-on or decorative assembly theme. They align teaching, relationships, school norms, family communication, leadership, and support systems with a broader developmental philosophy.

This distinction matters. A school can introduce isolated well-being exercises without changing anything meaningful about its culture. Students may be asked to practice gratitude while experiencing excessive pressure, weak belonging, punitive discipline, bullying, under-resourced classrooms, or teacher burnout. In those cases, positive education risks becoming superficial or even cynical. Positive education becomes more serious when it addresses how authority is exercised, how relationships are structured, how students are recognized, how setbacks are handled, and whether the institution itself supports belonging and growth.

Implementation is best understood across several levels:

Teacher development is especially important. Teachers cannot sustainably cultivate student well-being if they are unsupported, overburdened, undertrained, or working in unstable institutional conditions. Positive education requires adult capacity as well as student programming. If staff well-being is ignored, the approach becomes harder to sustain and easier to experience as another demand placed on already strained educators.

Implementation also requires developmental sensitivity. A well-being practice suitable for adolescents may not be appropriate for young children. A strengths exercise that feels empowering to one student may feel uncomfortable or culturally mismatched to another. A mindfulness exercise may be helpful for some students and unsettling for students with trauma histories if not handled carefully. Positive education should therefore be adaptive, not formulaic.

Back to top ↑

Scientific Evidence

A growing body of research suggests that school-based well-being programs can produce measurable benefits for some students and settings. Reviews of positive education interventions have linked such programs with improved emotional resilience, increased life satisfaction, stronger engagement with learning, and in some cases reductions in symptoms of anxiety or depression. The evidence is not uniform, and effects vary by design, duration, context, implementation quality, developmental stage, and population. But the field has moved well beyond speculation.

Programs that teach cognitive skills for interpreting adversity, regulating emotion, and responding more adaptively to difficulty have shown particular promise in some educational contexts. This is one reason resilience-oriented interventions became influential in the field’s development. More recent review work has also emphasized that the strongest positive education approaches are often embedded in whole-school culture rather than restricted to one-off classroom activities.

The evidence should still be interpreted carefully. Positive outcomes do not mean that every program works equally well, nor that school well-being interventions can compensate for poverty, trauma, underfunded schools, unstable housing, discrimination, disability exclusion, or high-stress institutional environments. The evidence supports cautious seriousness, not simplistic evangelism.

A useful way to read the evidence is to distinguish different kinds of outcomes:

Evidence quality also depends on study design. Randomized trials, longitudinal designs, implementation studies, mixed-method evaluation, and whole-school case studies each answer different questions. A randomized classroom intervention can help estimate program effects under controlled conditions, but it may not capture long-term culture change. A whole-school case study may reveal institutional dynamics but provide weaker causal inference. A serious evidence base needs both.

The most credible interpretation is therefore balanced: positive education has promising evidence, especially when well-being practices are developmentally appropriate and institutionally embedded. But it should not be sold as a universal cure. It is a framework for strengthening educational environments, not a shortcut around the hard work of school funding, teacher support, mental-health services, equity, disability inclusion, family engagement, and humane accountability systems.

Back to top ↑

School Climate, Belonging, and Institutional Design

Positive education becomes most powerful when it moves from individual intervention to institutional design. Schools are not neutral containers in which psychological skills are delivered. They are social systems with rules, hierarchies, routines, norms, expectations, relationships, histories, incentives, and forms of recognition. These systems shape student well-being every day.

Belonging is central. Students who feel seen, respected, and included are more likely to engage, ask for help, persist through difficulty, and imagine a future for themselves. Students who feel alienated, surveilled, stereotyped, excluded, or unsafe may disengage even when instruction is technically strong. Positive education therefore has to ask whether the school produces belonging structurally, not merely whether students are told to value it.

School climate includes more than friendliness. It includes whether discipline is fair, whether students have voice, whether teachers have time to know students well, whether bullying is taken seriously, whether disability accommodations are respected, whether cultural identity is honored, whether academic pressure is humane, whether mistakes are treated as part of learning, and whether adults model the forms of emotional and ethical development they expect from students.

Institutional design also matters for teachers. Teachers are often asked to support student well-being while navigating heavy workloads, emotional labor, administrative pressure, large class sizes, limited mental-health support, and accountability systems that reward narrow outputs. A positive education approach that ignores teacher well-being is incomplete. The adult ecology of the school shapes the student ecology.

The most serious versions of positive education therefore require systems thinking. They ask how curriculum, assessment, leadership, care structures, behavioral policies, teacher support, family engagement, and community partnerships interact. Well-being is not a single program inside the school. It is a property of the educational environment.

Back to top ↑

Education Policy and Global Adoption

Positive education has influenced policy discussions in several countries and educational systems. School leaders, universities, ministries, and international organizations increasingly discuss student well-being alongside academic metrics. This broader concern is visible in UNESCO’s current work on learner mental health and well-being, supportive learning environments, and the integration of health and well-being into education policy and planning.

These developments reflect a wider shift in educational thought. Schools are increasingly seen not only as sites of credentialing and skills acquisition, but as major institutions of human development. Questions of safety, psychological well-being, belonging, and social connection are no longer treated as peripheral to learning. They are increasingly understood as part of the conditions under which learning becomes possible and sustainable.

Positive education therefore sits at an important institutional boundary. It is not only a psychological project. It is also a policy and systems question about what educational institutions are for and how they should be organized. If well-being becomes part of educational policy, then governments and school systems must decide how it will be measured, funded, staffed, protected, and interpreted.

Policy adoption can create opportunity. It can normalize mental-health support, increase teacher training, broaden accountability systems, support whole-school climate work, and help schools recognize that academic success is connected to social and emotional development. But policy adoption can also create risk. If well-being becomes another metric in a high-pressure accountability regime, schools may feel compelled to produce positive data rather than improve conditions. Students may be surveyed repeatedly without meaningful support. Teachers may be asked to deliver well-being curricula without training, time, or institutional change.

A responsible policy approach should therefore include:

• Clear purpose. Well-being measurement should support student development, not institutional image management.
• Professional capacity. Teachers and staff need training, time, and support.
• Integrated services. Positive education should connect with counseling, mental-health referral systems, disability support, family engagement, and safeguarding.
• Equity attention. Data should be disaggregated responsibly to identify unequal burden without stigmatizing groups.
• Developmental appropriateness. Programs should fit age, context, culture, and student needs.
• Limits on high-stakes use. Well-being data should not become a tool for punitive ranking of students, teachers, or schools.

The policy promise of positive education lies in its ability to widen educational purpose. The policy danger lies in turning flourishing into another bureaucratic target. The difference depends on implementation, governance, and ethical restraint.

Back to top ↑

Equity, Culture, Disability, and Context

Positive education must be interpreted through equity. Students do not arrive at school with equal exposure to safety, health, housing stability, family resources, disability support, food security, community opportunity, or social recognition. Programs focused on optimism, resilience, gratitude, or strengths can be helpful, but they can also become harmful if they imply that students should adapt privately to conditions that institutions have a responsibility to change.

A student experiencing homelessness, racism, disability exclusion, family instability, food insecurity, bullying, grief, or trauma may need belonging, resilience, and emotional skills. But they also need material support, institutional protection, adult advocacy, safe environments, and fair access to learning. Positive education becomes credible when it pairs psychological development with structural realism.

Culture also matters. Ideas of happiness, autonomy, accomplishment, gratitude, purpose, and character vary across communities and traditions. A strengths activity may resonate differently across cultural contexts. A concept of purpose may be individual, familial, religious, civic, ecological, or communal. Emotional expression may be encouraged in one context and restrained in another. School belonging may be shaped by language, migration history, religion, race, gender, class, disability, and family expectations.

Disability inclusion is especially important. Positive education should not imply that flourishing requires a narrow model of emotional expression, social behavior, academic pace, or self-presentation. Neurodivergent students, disabled students, chronically ill students, and students with mental-health conditions may flourish in ways that require accommodation, flexibility, sensory safety, communication access, assistive technology, and recognition of different developmental trajectories. Well-being must not become conformity.

A more inclusive positive education framework asks:

• Who is being asked to become resilient, and why?
• Which students are already carrying disproportionate strain?
• Do well-being practices fit the cultural and developmental context?
• Are students with disabilities supported or simply expected to adapt?
• Does the school change its environment, or only train students to endure it?
• Are families and communities included in interpreting what flourishing means?

This equity lens does not weaken positive education. It strengthens it. It prevents the field from becoming a language of individual adjustment and keeps it connected to the broader conditions of educational flourishing.

Back to top ↑

Critiques and Limitations

Despite its promise, positive education has generated substantial debate among scholars and educators. One concern is contextual fit. Well-being programs must be adapted carefully to institutional, developmental, and cultural settings. A framework that works in one school system may not transfer seamlessly into another. This is especially important because many early positive psychology models were developed in relatively Western and affluent contexts.

A second concern is structural realism. Schools face constraints of time, funding, staff capacity, teacher training, accountability systems, family needs, student mental-health demand, and competing policy pressures. A school may value well-being and still lack the resources to implement sustained developmental programs. In such settings, positive education can become rhetorically attractive but practically thin.

A third concern is ideological overreach. Critics argue that focusing too exclusively on individual psychological skills may obscure structural factors such as inequality, family stress, institutional rigidity, racism, disability exclusion, or social disadvantage. A child’s well-being is not produced by mindset alone. If positive education is interpreted too narrowly, it risks becoming a language of adaptation rather than a broader effort to improve the conditions of educational life.

A fourth concern is measurement misuse. Well-being surveys can help schools identify needs, but they can also be misused if student data are collected without adequate privacy safeguards, if results are used for institutional branding, or if teachers and schools are ranked based on complex psychological outcomes they do not fully control. Well-being data require ethical governance.

A fifth concern is program superficiality. Short exercises can be useful, but they do not automatically change school culture. A gratitude exercise cannot repair chronic bullying. A strengths lesson cannot substitute for disability accommodations. A resilience curriculum cannot compensate for unsafe discipline practices or staff burnout. Positive education should not become a thin layer of positivity over unchanged institutional conditions.

These critiques do not invalidate the field. They define the conditions under which it can remain credible. Positive education is most defensible when it is evidence-informed, culturally responsive, developmentally appropriate, structurally aware, and embedded in real institutional change.

Back to top ↑

A Semi-Formal Framework for Positive Education

Positive education can be represented semi-formally as a joint developmental model in which academic and psychological outcomes interact across time. Let student flourishing at time \(t\) be represented as:

This formulation makes explicit what positive education often assumes: flourishing in schools depends on both psychological and institutional dimensions, not on either alone.

A dynamic model is also useful:

A dual-outcome model can represent the central educational claim:

A real-world implementation model must include constraints:

A distributional model is also necessary:

These equations are not meant to reduce education to formulas. They clarify the assumptions behind positive education: learning and well-being interact, school climate matters, strain matters, constraints matter, and equity matters.

Back to top ↑

Data Design and Measurement Notes

A serious positive education evaluation should measure more than whether students liked a program. It should include academic, psychological, relational, institutional, and equity-sensitive indicators. The goal is not to produce a single student happiness score. The goal is to understand the conditions under which students learn and flourish.

Several design principles follow:

• Measure school climate, not only student attitude. Student well-being should be interpreted in relation to institutional conditions.
• Disaggregate responsibly. Group patterns can reveal unequal burden, but they should not stigmatize students or communities.
• Protect privacy. Student well-being data require strong safeguards and clear limits on use.
• Use mixed methods. Surveys should be supplemented by interviews, student voice, teacher interpretation, and contextual evidence.
• Track implementation quality. Program effects depend on dosage, fidelity, teacher training, leadership support, and school culture.
• Avoid high-stakes misuse. Well-being scores should not be used to rank students, punish teachers, or brand schools.

Positive education measurement should support better environments. It should help schools ask what students need, where belonging is fragile, how stress is distributed, whether learning feels meaningful, and whether institutional practices support or undermine development.

Back to top ↑

R: Modeling Positive Education Outcomes

The following R workflow illustrates how a researcher might model positive education outcomes using repeated student-level observations that include both academic and well-being variables. The example estimates a school flourishing outcome while preserving the distinction between academic, relational, psychological, climate, and strain contributors.

# Positive education modeling workflow
#
# Purpose:
#   Estimate school flourishing using repeated student observations
#   that include academic, psychological, relational, climate, and strain
#   indicators.
#
# Notes:
#   This workflow is for research, teaching, and exploratory analysis.
#   It is not a clinical, diagnostic, therapeutic, workplace-screening,
#   employment-selection, disciplinary, or individual student assessment tool.

library(tidyverse)
library(psych)
library(lme4)
library(lmerTest)
library(broom.mixed)
library(emmeans)
library(performance)

# Expected columns:
# student_id, school_id, wave,
# academic_score, engagement, belonging, resilience,
# life_satisfaction, school_climate, purpose_learning,
# teacher_support, stress_load, exclusion_exposure

df <- read_csv("data/positive_education_panel.csv")

panel <- df %>%
  mutate(
    student_id = as.factor(student_id),
    school_id = as.factor(school_id),
    wave = as.integer(wave)
  ) %>%
  filter(complete.cases(
    academic_score,
    engagement,
    belonging,
    resilience,
    life_satisfaction,
    school_climate,
    purpose_learning,
    teacher_support,
    stress_load,
    exclusion_exposure
  ))

# Reliability check for school flourishing indicators.
school_items <- panel %>%
  select(
    engagement,
    belonging,
    resilience,
    life_satisfaction,
    purpose_learning,
    teacher_support
  )

psych::alpha(school_items)

panel_scored <- panel %>%
  mutate(
    academic_z = as.numeric(scale(academic_score)),
    engagement_z = as.numeric(scale(engagement)),
    belonging_z = as.numeric(scale(belonging)),
    resilience_z = as.numeric(scale(resilience)),
    life_satisfaction_z = as.numeric(scale(life_satisfaction)),
    climate_z = as.numeric(scale(school_climate)),
    purpose_z = as.numeric(scale(purpose_learning)),
    teacher_support_z = as.numeric(scale(teacher_support)),
    stress_z = as.numeric(scale(stress_load)),
    exclusion_z = as.numeric(scale(exclusion_exposure)),
    relational_support = rowMeans(
      select(., belonging_z, teacher_support_z),
      na.rm = TRUE
    ),
    psychological_resources = rowMeans(
      select(., resilience_z, life_satisfaction_z, purpose_z),
      na.rm = TRUE
    ),
    school_flourishing =
      0.25 * academic_z +
      0.20 * engagement_z +
      0.20 * relational_support +
      0.20 * psychological_resources +
      0.20 * climate_z -
      0.15 * stress_z -
      0.15 * exclusion_z,
    wave_c = as.numeric(scale(wave, center = TRUE, scale = FALSE)),
    academic_c = as.numeric(scale(academic_z, center = TRUE, scale = FALSE)),
    climate_c = as.numeric(scale(climate_z, center = TRUE, scale = FALSE)),
    belonging_c = as.numeric(scale(belonging_z, center = TRUE, scale = FALSE)),
    resilience_c = as.numeric(scale(resilience_z, center = TRUE, scale = FALSE)),
    support_c = as.numeric(scale(teacher_support_z, center = TRUE, scale = FALSE)),
    stress_c = as.numeric(scale(stress_z, center = TRUE, scale = FALSE)),
    exclusion_c = as.numeric(scale(exclusion_z, center = TRUE, scale = FALSE))
  )

model_school <- lmer(
  school_flourishing ~
    wave_c +
    academic_c +
    climate_c +
    belonging_c +
    resilience_c +
    support_c -
    stress_c -
    exclusion_c +
    academic_c:climate_c +
    belonging_c:support_c +
    resilience_c:stress_c +
    (1 + wave_c | student_id) +
    (1 | school_id),
  data = panel_scored,
  REML = FALSE
)

summary(model_school)
performance::check_model(model_school)

emm_academic_climate <- emmeans(
  model_school,
  ~ academic_c | climate_c,
  at = list(
    academic_c = c(-1, 0, 1),
    climate_c = c(-1, 0, 1),
    belonging_c = 0,
    resilience_c = 0,
    support_c = 0,
    stress_c = 0,
    exclusion_c = 0,
    wave_c = 0
  )
)

emm_resilience_stress <- emmeans(
  model_school,
  ~ resilience_c | stress_c,
  at = list(
    resilience_c = c(-1, 0, 1),
    stress_c = c(-1, 0, 1),
    academic_c = 0,
    climate_c = 0,
    belonging_c = 0,
    support_c = 0,
    exclusion_c = 0,
    wave_c = 0
  )
)

dir.create("outputs", showWarnings = FALSE)

write_csv(
  broom.mixed::tidy(model_school, effects = "fixed", conf.int = TRUE),
  "outputs/positive_education_model_results.csv"
)

write_csv(
  broom.mixed::tidy(model_school, effects = "ran_pars", conf.int = TRUE),
  "outputs/positive_education_random_effects.csv"
)

write_csv(
  as.data.frame(emm_academic_climate),
  "outputs/academic_by_climate_estimated_margins.csv"
)

write_csv(
  as.data.frame(emm_resilience_stress),
  "outputs/resilience_by_stress_estimated_margins.csv"
)

write_csv(
  panel_scored,
  "outputs/positive_education_scored_panel.csv"
)

school_summary <- panel_scored %>%
  group_by(school_id) %>%
  summarize(
    mean_school_flourishing = mean(school_flourishing, na.rm = TRUE),
    mean_academic = mean(academic_z, na.rm = TRUE),
    mean_engagement = mean(engagement_z, na.rm = TRUE),
    mean_belonging = mean(belonging_z, na.rm = TRUE),
    mean_climate = mean(climate_z, na.rm = TRUE),
    mean_stress = mean(stress_z, na.rm = TRUE),
    mean_exclusion = mean(exclusion_z, na.rm = TRUE),
    .groups = "drop"
  ) %>%
  arrange(desc(mean_school_flourishing))

write_csv(
  school_summary,
  "outputs/positive_education_school_summary.csv"
)

This workflow is useful because it treats positive education as a joint developmental system rather than as a simple add-on to academic instruction. It allows researchers to test how climate, belonging, academic competence, teacher support, exclusion, resilience, and stress interact in shaping school flourishing.

The resilience-by-stress interaction is especially important. If resilience appears protective only under moderate stress but not under high strain, that finding would support a structural interpretation: students need supportive environments, not merely coping skills. Likewise, an academic-by-climate interaction can show whether academic growth is more strongly associated with flourishing in supportive school climates.

Back to top ↑

Python: Network Analysis of School Well-Being Systems

The following Python example models positive education as a connected system of school-based variables rather than as a single intervention score. It estimates a sparse partial-correlation network across academic performance, engagement, belonging, resilience, life satisfaction, school climate, teacher support, purpose, stress, and exclusion exposure.

"""
Positive education network workflow

Purpose:
    Estimate a sparse network of school well-being indicators using
    partial correlations, then summarize centrality and edge structure.

Use:
    Research, teaching, exploratory systems analysis, and school-climate
    measurement design.

Not for:
    Clinical diagnosis, therapeutic decision-making, employment selection,
    disciplinary action, student ranking, or individual well-being assessment.
"""

from pathlib import Path

import matplotlib.pyplot as plt
import networkx as nx
import numpy as np
import pandas as pd

from sklearn.covariance import GraphicalLassoCV
from sklearn.impute import SimpleImputer
from sklearn.preprocessing import StandardScaler

DATA_PATH = Path("data/positive_education_network.csv")
OUTPUT_DIR = Path("outputs")
OUTPUT_DIR.mkdir(exist_ok=True)

cols = [
    "academic_score",
    "engagement",
    "belonging",
    "resilience",
    "life_satisfaction",
    "school_climate",
    "teacher_support",
    "purpose_learning",
    "stress_load",
    "exclusion_exposure",
]

df = pd.read_csv(DATA_PATH)

missing_cols = [col for col in cols if col not in df.columns]
if missing_cols:
    raise ValueError(f"Missing expected columns: {missing_cols}")

imputer = SimpleImputer(strategy="median")
X = pd.DataFrame(imputer.fit_transform(df[cols]), columns=cols)

scaler = StandardScaler()
X_scaled = pd.DataFrame(scaler.fit_transform(X), columns=cols)

X_scaled["school_flourishing_index"] = (
    0.25 * X_scaled["academic_score"] +
    0.20 * X_scaled["engagement"] +
    0.20 * X_scaled["belonging"] +
    0.20 * X_scaled["resilience"] +
    0.20 * X_scaled["life_satisfaction"] +
    0.20 * X_scaled["school_climate"] +
    0.20 * X_scaled["teacher_support"] +
    0.15 * X_scaled["purpose_learning"] -
    0.15 * X_scaled["stress_load"] -
    0.15 * X_scaled["exclusion_exposure"]
)

glasso = GraphicalLassoCV()
glasso.fit(X_scaled[cols])

precision = glasso.precision_
partial_corr = -precision / np.sqrt(np.outer(np.diag(precision), np.diag(precision)))
np.fill_diagonal(partial_corr, 0)

partial_df = pd.DataFrame(partial_corr, index=cols, columns=cols)

threshold = 0.08
G = nx.Graph()

for node in cols:
    G.add_node(node)

for i, source in enumerate(cols):
    for j, target in enumerate(cols):
        if j > i:
            weight = partial_df.iloc[i, j]
            if abs(weight) >= threshold:
                G.add_edge(source, target, weight=weight, sign=np.sign(weight))

degree = nx.degree_centrality(G)
betweenness = nx.betweenness_centrality(G, weight="weight")

try:
    eigenvector = nx.eigenvector_centrality_numpy(G, weight="weight")
except nx.NetworkXException:
    eigenvector = {node: np.nan for node in G.nodes()}

centrality = pd.DataFrame({
    "node": list(G.nodes()),
    "degree_centrality": [degree[node] for node in G.nodes()],
    "betweenness_centrality": [betweenness[node] for node in G.nodes()],
    "eigenvector_centrality": [eigenvector[node] for node in G.nodes()],
}).sort_values(
    ["eigenvector_centrality", "degree_centrality"],
    ascending=False
)

edge_table = pd.DataFrame([
    {
        "source": source,
        "target": target,
        "partial_correlation": data["weight"],
        "absolute_weight": abs(data["weight"]),
        "sign": "positive" if data["weight"] > 0 else "negative",
    }
    for source, target, data in G.edges(data=True)
]).sort_values("absolute_weight", ascending=False)

centrality.to_csv(OUTPUT_DIR / "positive_education_network_centrality.csv", index=False)
edge_table.to_csv(OUTPUT_DIR / "positive_education_network_edges.csv", index=False)
partial_df.to_csv(OUTPUT_DIR / "positive_education_partial_correlations.csv")
X_scaled.to_csv(OUTPUT_DIR / "positive_education_scaled_indices.csv", index=False)

print("\nCentrality summary:")
print(centrality)

print("\nStrongest edges:")
print(edge_table.head(15))

plt.figure(figsize=(12, 9))
pos = nx.spring_layout(G, seed=42, k=0.85)

positive_edges = [(u, v) for u, v in G.edges() if G[u][v]["weight"] > 0]
negative_edges = [(u, v) for u, v in G.edges() if G[u][v]["weight"] < 0]

nx.draw_networkx_nodes(G, pos, node_size=1800)
nx.draw_networkx_labels(G, pos, font_size=9)

nx.draw_networkx_edges(
    G,
    pos,
    edgelist=positive_edges,
    width=[abs(G[u][v]["weight"]) * 5 for u, v in positive_edges],
    alpha=0.75,
)

nx.draw_networkx_edges(
    G,
    pos,
    edgelist=negative_edges,
    width=[abs(G[u][v]["weight"]) * 5 for u, v in negative_edges],
    style="dashed",
    alpha=0.75,
)

plt.title("Partial Correlation Network of Positive Education Outcomes")
plt.axis("off")
plt.tight_layout()
plt.savefig(OUTPUT_DIR / "positive_education_network.png", dpi=300)
plt.close()

This type of analysis can reveal whether belonging, school climate, teacher support, resilience, or academic engagement functions as the more central leverage point in a given school system. That matters because positive education is often more effective when it changes structurally central features of school life rather than layering isolated exercises onto unchanged institutions.

Network models should not be interpreted as causal proof. They are exploratory tools for identifying relationships that may deserve further investigation, longitudinal testing, qualitative interpretation, and school-community review. If belonging appears central, the next step is not simply to tell students to belong more. It is to examine the institutional conditions that produce or block belonging.

Back to top ↑

Interpretation and Responsible Use

Because positive education involves children and adolescents, responsible use is especially important. School well-being data can be sensitive even when it is not clinical. A student’s responses about belonging, stress, life satisfaction, loneliness, resilience, or school climate may reveal vulnerability, exclusion, family strain, disability-related barriers, trauma exposure, or mental-health concerns. These data should be handled with care.

The code examples above are designed for population-level research, teaching, exploratory modeling, and school-climate analysis. They should not be used as clinical diagnostic instruments, therapeutic decision tools, disciplinary tools, student ranking systems, teacher ranking systems, employment-selection tools, or individual student well-being assessments.

Several principles follow:

• Do not collapse students into scores. Well-being indicators are partial signals, not full accounts of a child’s life.
• Measure school conditions as well as student attitudes. Belonging, stress, and resilience should be interpreted in relation to climate, safety, support, and institutional practice.
• Avoid moralizing distress. Low well-being may reflect grief, exclusion, trauma, disability barriers, poverty, bullying, discrimination, or institutional strain.
• Protect privacy. Student data require clear consent procedures, aggregation rules, access limits, and careful communication.
• Use data to improve conditions. Measurement should lead to support, not surveillance or public relations.
• Include student voice. Quantitative results should be interpreted alongside lived experience and student perspectives.
• Respect developmental and cultural context. Well-being practices should be adapted to age, culture, language, disability, and community realities.

The purpose of positive education measurement should be institutional learning. Schools should use evidence to ask where students feel supported, where they feel burdened, where relationships are fragile, where staff need support, and where climate change is necessary. Data should help improve educational environments, not make students responsible for adapting to harmful ones.

Back to top ↑

GitHub Repository

The companion repository for this article organizes the R, Python, data-schema, and documentation materials into a reproducible workflow for positive education analysis. It includes sample data dictionaries, scripts for model estimation, network-analysis outputs, validation notes, and guidance for responsible interpretation.

Back to top ↑

Conclusion

Positive education represents one of the most significant efforts to apply the science of well-being within educational institutions. By combining academic instruction with psychological resources such as resilience, belonging, purpose, emotional awareness, character, and supportive relationships, the field seeks to prepare students not only for academic success but for meaningful and flourishing lives.

Its strongest contribution is not the promise of cheerful classrooms. It is the claim that schooling should be judged partly by whether it supports the conditions under which young people can genuinely develop. This means that positive education must be more than a toolkit of exercises. It must become a serious framework for thinking about curriculum, relationships, school climate, teacher support, student voice, disability inclusion, equity, and institutional responsibility.

The field remains promising, but its promise depends on humility. Positive education should not ask students to adapt privately to preventable harm. It should not reduce well-being to positivity. It should not allow schools to collect well-being data without improving conditions. And it should not become a substitute for adequate funding, mental-health support, safe learning environments, or social justice.

At its best, positive education widens the moral and developmental imagination of schooling. It reminds educators and policymakers that students are not test-score machines or future workers only. They are developing human beings whose learning is inseparable from belonging, meaning, dignity, care, and the possibility of becoming capable participants in shared life.

Back to top ↑

Related articles

• Positive Psychology article map
• The PERMA Model of Well-Being
• Self-Determination Theory and Positive Psychology
• Meaning and Purpose in Positive Psychology
• Hope Theory in Positive Psychology
• Explanatory Style and Optimism
• Character Strengths and Virtues in Positive Psychology
• The Three Good Things Exercise
• Positive Psychology Interventions
• Positive Psychology and Public Health

Further reading

• Seligman, M.E.P., Ernst, R.M., Gillham, J., Reivich, K. and Linkins, M. (2009) ‘Positive education: Positive psychology and classroom interventions’, Oxford Review of Education, 35(3), pp. 293–311. Available at: https://ppc.sas.upenn.edu/sites/default/files/positiveeducationarticle2009.pdf.
• Waters, L. (2011) ‘A review of school-based positive psychology interventions’, The Australian Educational and Developmental Psychologist, 28(2), pp. 75–90.
• Waters, L., White, M.A. and Murray, S. (2022) ‘Positive education: Theory, practice, and evidence’, Frontiers in Psychology, 13, 908504. Available at: https://www.frontiersin.org/articles/10.3389/fpsyg.2022.908504/full.
• UNESCO (2025) Education for health and well-being. Available at: https://www.unesco.org/en/health-education/need-know.

Back to top ↑

References

• Positive Psychology Center (n.d.) Resilience Training for Schools. Available at: https://ppc.sas.upenn.edu/services/resilience-training-schools.
• Positive Psychology Center (n.d.) Penn Resilience Program and PERMA Workshops. Available at: https://ppc.sas.upenn.edu/services/penn-resilience-training.
• Seligman, M.E.P., Ernst, R.M., Gillham, J., Reivich, K. and Linkins, M. (2009) ‘Positive education: Positive psychology and classroom interventions’, Oxford Review of Education, 35(3), pp. 293–311. Available at: https://ppc.sas.upenn.edu/sites/default/files/positiveeducationarticle2009.pdf.
• UNESCO (2025) Education for health and well-being. Available at: https://www.unesco.org/en/health-education/need-know.
• UNESCO (2025) Wellbeing and mental health at school: Guidelines for policymakers. Available at: https://healtheducationresources.unesco.org/library/documents/wellbeing-and-mental-health-school-guidelines-education-policymakers.
• Waters, L. (2011) ‘A review of school-based positive psychology interventions’, The Australian Educational and Developmental Psychologist, 28(2), pp. 75–90.
• Waters, L., White, M.A. and Murray, S. (2022) ‘Positive education: Theory, practice, and evidence’, Frontiers in Psychology, 13, 908504. Available at: https://www.frontiersin.org/articles/10.3389/fpsyg.2022.908504/full.

Back to top ↑