Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Community Prevention and Promotion Systems

Abstract. Community prevention and promotion systems coordinate population-level risk reduction, protective environments, collective capacity, institutional implementation, and sustained psychological well-being. This fundamental research monograph develops a critical narrative synthesis of ecological community psychology, prevention science, developmental intervention, health promotion, participatory governance, causal evaluation, and implementation research. It examines the historical formation of risk and protective factor systems, empowerment, universal and selective prevention, and the distinction between preventing adverse outcomes and promoting positive functioning. Original community- and school-randomized evidence is assessed through Communities That Care, PROSPER, Project Northland, the Community Intervention Trial for Smoking Cessation, and the Good Behavior Game, with systematic reviews of social-emotional learning and community engagement. Findings demonstrate condition-specific benefits alongside important null, mixed, and time-dependent results. The synthesis identifies the necessity of accounting for cluster-level assignment, implementation fidelity, population reach, equity, developmentally changing exposure, and long-term maintenance. It develops a layered framework for evaluating systems and identifies unresolved questions about feedback, spillovers, adaptation, costs, and emerging Human–Artificial institutional interactions. The work is a critical narrative scientific synthesis, rather than an original trial or registered meta-analysis.

Keywords: community prevention; health promotion; community psychology; universal prevention; selective prevention; indicated prevention; risk and protective factors; collective capacity; community empowerment; community interventions; Communities That Care; PROSPER; Project Northland; COMMIT; Good Behavior Game; cluster-randomized trials; collective efficacy; implementation science; participatory research; RE-AIM; health equity; program sustainability; longitudinal evaluation; Artificial Era

The Research Object and the Unit of Prevention

Community prevention and promotion systems are organized arrangements of people, institutions, decision rules, resources, and interventions designed to alter the distribution of risks and protective conditions across a defined population. A community may be geographically delimited, institutionally organized, or connected through sustained relationships; the scientific unit must be specified before outcomes are measured. Prevention concerns reducing the incidence, onset, severity, or progression of adverse outcomes, whereas promotion concerns developing capacities, conditions, and opportunities that support functioning and well-being. The two functions frequently interact because protective relationships, supportive norms, and institutional resources can influence several outcomes simultaneously. The historical World Health Organization (1986) Charter defined health promotion through increased control over health and explicitly assigned priority to collective action, supportive environments, and community capacity.

A system-level intervention has multiple causal layers: who sets priorities, which conditions are selected for change, how programs are implemented, which institutions maintain them, and how exposure accumulates through time. Its outcomes include population incidence, differential exposure, participation, and durability, rather than only an average symptom score among people who attended a session. An effective course delivered to a few volunteers can have high individual efficacy with limited population reach; a widespread policy can influence community outcomes even when its individual behavioral effect is modest. Glasgow et al. (1999) supplied a practical framework for distinguishing reach, effectiveness, adoption, implementation, and maintenance in evaluations.

The central research question is whether organized community action changes psychologically consequential conditions through identifiable mechanisms. This requires distinguishing the intervention package, the implementation system, and the ecology into which both are introduced. The dossier integrates original prevention trials, community-psychological theory, historical public-health documents, process research, and structured reviews. It is a critical narrative synthesis, with explicitly identified source limitations; no original field trial, registered systematic review, or statistical meta-analysis was conducted for this monograph.

Historical Formation of Community Prevention Science

The scientific history of prevention systems reaches across public health, social reform, education, community mental health, developmental psychology, and epidemiology. The transition from treating disorders after onset toward altering their antecedents acquired methodological force as researchers began to document risk exposure across cohorts and settings. In community psychology, Kelly (1966) drew attention to ecological constraints on mental health services. That article is a historically significant statement of the field's ecological direction, although its indexed record alone does not warrant claims about precise empirical treatment effects. Community prevention subsequently developed around collaborations among schools, families, neighborhood organizations, health agencies, and local authorities.

The World Health Organization (1986) Charter consolidated an international health-promotion language in which healthy public policy, supportive environments, community action, personal skills, and redirected services are mutually reinforcing priorities. Its normative framework can organize intervention design, but the Charter itself is not an experimental demonstration that every multisector intervention works. In the same era, Rappaport (1987) argued that empowerment concerns the phenomena community psychology should understand, emphasizing people and institutions exercising meaningful control over their circumstances. Prevention thus acquired a dual research agenda: identifying modifiable antecedents of harms and examining who has authority over the preventive system.

Large controlled community programs revealed both promise and difficulty. The COMMIT smoking trials generated important null and mixed results COMMIT Research Group (1995a) COMMIT Research Group (1995b). Project Northland tested school, family, peer, and community components in a geographically organized adolescent cohort Perry et al. (1996). Communities That Care later tested a coalition-led approach in randomized towns Hawkins et al. (2014). This sequence reflects an improving capacity to distinguish adoption of prevention activities, verified behavioral outcomes, and later maintenance. Evidence accrued unevenly across intervention types, with stronger support for some particular bundles than for the idea that coalition formation itself necessarily changes health.

Prevention, Promotion, and the Risk–Protection Framework

Risk factors are measurable characteristics of individuals, relationships, organizations, or environments that predict increased probability of a defined adverse outcome in a specified population. Protective factors are characteristics associated with lower probability or with attenuation of particular risks. A factor's predictive association does not itself prove that changing it will change an outcome: it can serve as a proxy for structural disadvantage, reflect prior behavior, or depend on circumstances outside a program's control. The foundational synthesis by Hawkins et al. (1992) reviewed risk and protective factors for adolescent substance problems and called for prevention strategies built upon identification, targeted response, and controlled evaluation.

Prevention can be universal, directed toward everyone in a defined population; selective, directed toward subgroups with elevated statistical risk; or indicated, directed toward people showing early signs that have not yet become the target disorder. These classifications describe the intended exposure population, not a hierarchy of moral merit. A universal program may protect people who would never have met selection thresholds, while a selective strategy can concentrate resources where conditions are most adverse. The optimal allocation depends on baseline incidence, treatment effectiveness, costs, uptake, potential stigma, and unintended effects. Promotion initiatives may likewise be broad or targeted, depending on whether the objective is population-wide support or reduction of an inequity.

The ecological interpretation of risk locates processes in developmentally and institutionally structured contexts. Parenting opportunities, neighborhood safety, school connectedness, financial precarity, discrimination, and access to care may be correlated; changing one variable while leaving related conditions unchanged can weaken the predicted effect. Biglan et al. (2012) proposed nurturing environments as an organizing concept for behavioral and mental-health prevention, suggesting that recurring environmental conditions shape multiple outcomes. That synthesis is theoretically useful, but specific causal claims still require interventions or credible natural experiments that identify the effect of altering those conditions.

Ecological Levels, Settings, and Interdependence

Individuals encounter prevention systems through households, peer relationships, schools, workplaces, neighborhoods, organizations, and digital communication environments. Each setting has characteristic exposures, rules, expectations, and access barriers. Interdependence means a change in one setting can affect processes in another: a school program may modify family communication, a transportation policy may change access to mental-health services, and a workplace intervention may alter stress experienced at home. The importance of a setting follows from repeated contact and available mechanisms, not simply its physical size. WHO's healthy-settings approach World Health Organization (n.d.) describes a whole-system strategy integrating environmental, organizational, and personal determinants.

Ecological explanation demands attention to the boundary chosen for analysis. A school district may contain students from multiple neighborhoods, and residents may receive care outside the town where they sleep. The population exposed to the intervention may differ from the population used to calculate its outcome. This can produce ecological dilution or apparent impact unrelated to the intervention. Cross-boundary spillovers become particularly important when neighboring communities share providers, media markets, school networks, or social platforms. Systems cannot always be neatly randomized at the level where behavior is ultimately produced.

The empirical finding by Sampson et al. (1997) linking neighborhood collective efficacy with differences in violent crime illustrates the relevance of relational environments. Its multilevel study surveyed 8,782 Chicago residents in 343 neighborhoods and found an association after multiple controls. Because this was an observational investigation, the association should inform hypotheses about social cohesion and willingness to intervene, rather than be treated as experimental proof that a coalition program will reduce violence. The distinction between an ecological correlate and an experimentally manipulated component is indispensable to a scientifically coherent community prevention architecture.

Research Questions, Evidence Sources, and Interpretive Rules

This monograph asks how community prevention systems select risks, organize protective conditions, recruit collective action, deliver effective interventions, adapt to local context, and maintain outcomes beyond temporary external support. It also asks which designs distinguish system effects from secular trends, how equity is assessed, and when promotion changes the underlying distribution of opportunity. Foundational theory is read as conceptual evidence about what researchers proposed; randomized intervention studies provide narrower causal evidence for their tested packages; systematic reviews aggregate effects under specified inclusion criteria; implementation studies examine why a practice can fail to reach its target population.

The source selection used direct inspection of publicly accessible study records and, where available, open full texts and authoritative organizational documents. Papers were selected for their direct relevance to community prevention, clear empirical design, historical importance, contrary results, and methodological influence. The resulting collection is purposive, not exhaustive. It is unsuitable for calculating a pooled mean across trials, claiming formal PRISMA compliance, or estimating the proportion of all published studies that yielded positive findings. A bibliographic record containing an abstract supports claims within that abstract, not uninspected subgroup analyses or supplementary tables.

Every numerical study result is attributed to the source that reports it. Interpretations distinguish between initiation and prevalence, individual attendance and community reach, implementation fidelity and intervention effectiveness, association and causality, and statistical significance and practical consequence. Negative or mixed evidence is analyzed together with successful programs. Independent replication, adequate site numbers, measurement invariance, and long follow-up are treated as continuing research requirements, especially when programs move from tightly supported trials to routine institutional operations.

Collective Capacity, Empowerment, and Decision Authority

A coalition is an organizational arrangement, while collective capacity is the ability to recognize problems, coordinate action, mobilize relevant resources, and sustain decisions. Capacity can include trust, representation, institutional access, technical assistance, organizational competence, and authority to redirect funding. These dimensions should be measured separately. A crowded meeting demonstrates participation, but it does not establish that participants control program selection or budget allocation. Rappaport (1987) framed empowerment as a core concern of community psychology, inviting researchers to ask who controls both decisions and definitions of success.

Community governance typically joins residents, practitioners, schools, service agencies, and funders whose interests overlap imperfectly. Governance determines who can define target outcomes and which risks become visible. Coalitions may attract highly resourced organizations while excluding people experiencing the greatest exposure; this can yield programs optimized for those easiest to reach. A credible assessment therefore records membership composition, decision rights, leadership turnover, meeting access, budget control, and whether marginalized groups can reject or revise an intervention. These are possible mechanisms of implementation and legitimacy, rather than substitutes for outcome evaluation.

The causal role of collective capacity can be investigated through longitudinal mediation analysis, randomized support for coalition processes, comparative case studies, and documented changes in decision authority. Yet capacity can improve while health outcomes remain unchanged, especially when funding, housing, discrimination, or service capacity place stronger constraints on behavior. Program evaluators should therefore maintain distinct measures of organizational development, equity of influence, intervention uptake, and population health. The positive value of participation must be analyzed alongside its actual distribution and effects.

Protective Relationships and Nurturing Environments

Relationships may protect psychological functioning through reliable care, behavioral modeling, emotional co-regulation, practical assistance, social expectations, and opportunities for learning. Their effects depend on relational quality, developmental stage, and context. Protective relationships can be produced through family support, peer mentoring, school climate, neighborhood participation, and community institutions. The nurturing-environments account developed by Biglan et al. (2012) treats reinforcement of prosocial behavior and reduction of coercive interactions as processes relevant to multiple behavioral and mental-health outcomes.

A prevention system must distinguish between the proximal mechanism and the organization used to deliver it. Teaching parenting skills is one route to changing family interaction, whereas providing reliable childcare, transportation, food security, or safe public environments may change the conditions under which those skills can be practiced. A short training program can demonstrate knowledge gain without demonstrating that protective relationships have improved. For interpretation, researchers need measures of experienced relationships, repeated exposure, social connection, and later functioning, preferably observed from multiple respondents or independent records.

Promotion aims can be framed positively: greater belonging, agency, coping resources, emotional literacy, meaningful roles, and safe participation. These outcomes may be scientifically valuable even in cohorts with low disorder prevalence. Catalano et al. (2004) reviewed positive youth development programs as an approach to examining developmental assets and program outcomes. At the population level, the question extends beyond whether individuals acquired a skill to whether institutions reliably provide opportunities to use it. Sustainable promotion therefore combines human learning with the organizational conditions that make learning consequential.

Social Norms, Behavioral Contagion, and Reinforcement

Community prevention can modify behavioral expectations through peer networks, school norms, availability of substances, shared sanctions, and publicly observable models of conduct. Normative change may alter perceived acceptability, anticipated consequences, or the repertoire of alternatives accessible to members. It can also create unintended pressures when programs stigmatize groups defined as risky. A scientifically defensible norm intervention specifies the pathway from a change in social information or incentives to an observable outcome, and measures whether that pathway operates at the intended level.

The Baltimore Good Behavior Game provides an important illustration of a population-exposure mechanism embedded in a routine classroom. In a trial reported by Kellam et al. (2008), first- and second-grade classrooms received behavior management intended to reduce early aggressive and disruptive behavior. Follow-up into young adulthood indicated some favorable outcomes, notably for males who displayed greater early behavioral risk, while a later cohort experienced weaker mentoring and monitoring and generally smaller effects. This history connects early developmental conditions with long-term prevention while emphasizing that the implementation context matters.

Norms transmitted through media and digital communities may span geographical boundaries, creating spillovers across intervention and comparison areas. Evaluations of norm change should measure actual exposure and whether participants accurately perceive peer behavior. Self-reported normative beliefs can be affected by social desirability and program messaging. When network-level influence is suspected, the research design should explicitly anticipate interference among members; individual randomization that assumes isolated responses may fail to identify the effect of a network intervention.

Social Development, Timing, and Accumulated Exposure

Development alters both the prevalence of risks and the processes available for intervention. The transition into adolescence increases opportunities for peer influence, autonomy, and exposure to new environments; a prevention package effective in middle school may require redesign as youth approach graduation. The same behavioral outcome can have different developmental antecedents at ages ten and eighteen. Programs must therefore specify the targeted developmental window, expected latency of effects, and outcome horizon.

The trajectory of Project Northland demonstrates this issue. Its early school, family, peer, and community components were associated with lower reported onset and prevalence of adolescent alcohol use Perry et al. (1996). Its later analysis Perry et al. (2002) found especially favorable results during early adolescence, described a weakening during a relatively minimal intervention interval, and reported mixed later results. These observations suggest that maintenance over successive developmental transitions is a substantive intervention question, not a routine administrative detail.

Cumulative prevention effects can occur through repeated reduction of exposure or delay of initiation, but estimates depend on the outcome definition. A cumulative measure of whether someone has ever initiated a behavior cannot decrease once initiation occurs; a current prevalence measure can change in either direction. Analyses of onset should model time-to-event or incidence, with censoring and attrition handled explicitly, whereas assessments of symptom burden and current behavior require repeated measures. The distinction becomes decisive when comparing late adolescent results of a multiyear intervention.

Institutional, Material, and Policy Mechanisms

Community systems can influence psychological outcomes by shaping access to resources and the regularity of daily life. Public transit, housing security, recreational spaces, safe schooling, financial support, service availability, and anti-discrimination rules alter exposure and choice. A theory focused solely on individual motivation can overlook circumstances that make a behavior difficult or costly. The settings-based perspective World Health Organization (n.d.) connects environmental, organizational, and personal determinants and invites interventions that operate through multiple channels.

Policies change incentives and availability across entire populations, sometimes with little direct contact between a resident and an intervention worker. Their effects can be evaluated through interrupted time series, staggered adoption designs, comparative population panels, or natural experiments, provided assumptions about trends and selection are examined. Such evidence may offer greater external reach than tightly supervised programs, but simultaneous policy changes and migration can complicate causal attribution. No single aggregate before–after difference proves a policy effect.

Material supports and psychosocial programs can complement one another. A family-skills intervention may gain reach when schedules and transport barriers are reduced; a community media campaign may be less consequential when services are inaccessible. The most compelling systems explanations identify these interactions and predict which component will become ineffective if a complementary resource disappears. Treating complexity as a model of interacting mechanisms allows specific hypotheses to be tested, rather than treating complexity as a reason to dispense with causal evidence.

Communities That Care: Coalition-Guided Prevention Trial

Communities That Care (CTC) is a community prevention operating system that convenes local coalitions, examines epidemiologic risk and protection profiles, selects tested programs, and organizes implementation. Its scientific importance lies in testing a decision-and-delivery system rather than a single curriculum. The Community Youth Development Study randomized 24 small towns in seven U.S. states within matched groups; a cohort of 4,407 students who entered fifth grade was followed over multiple years. In the intervention towns, coalitions received training and assistance and implemented evidence-based activities for youth, families, and schools Hawkins et al. (2014).

The intermediate results Hawkins et al. (2012) reported lower average targeted-risk trajectories and favorable differences in several onset or prevalence outcomes by tenth grade, after the period of initial installation. By twelfth grade, Hawkins et al. (2014) reported greater sustained abstinence from any drug use (adjusted risk ratio 1.32, 95% CI 1.06–1.63), alcohol use (1.31, 1.09–1.58), and cigarette smoking (1.13, 1.01–1.27); cumulative violent behavior was also lower (adjusted risk ratio 0.86, 95% CI 0.76–0.98). At the final follow-up, differences were not significant for the measured current prevalence of substance use, recent delinquency or violence, or targeted risk. The same trial therefore supports a meaningful, outcome-specific prevention effect and a clear boundary on claims about persistent current behavior.

These are community-randomized estimates from a particular U.S. small-town design with extensive implementation support; they do not establish identical effects in cities, different political systems, or settings with weaker services. The study's long observation and distinction among initiation, prevalence, and risk are notable strengths. Interpretation should respect the number of randomized clusters as well as the number of surveyed adolescents. The principal source article disclosed a relevant program-distribution relationship involving one coauthor Hawkins et al. (2014), reinforcing the need for transparent independence and replication.

PROSPER and the Community–University Delivery System

The PROSPER research program tested a delivery infrastructure linking local schools, community teams, and university-supported expertise to implement family- and school-based evidence-based preventive programs. In the cluster-randomized longer-term study Spoth et al. (2013), 28 public school districts were assigned to a partnership system or usual programming, with a baseline cohort of 11,960 students. The scientific object is simultaneously the tested prevention packages and the local system that selected and sustained their delivery.

PROSPER evidence shows why evaluation of service infrastructure is necessary: a strong curriculum cannot influence a student who is not reached, while training, recruitment, administrative support, and follow-up may determine exposure. The long-term paper assessed substance misuse approximately 6.5 years after baseline, retaining the distinction between tested delivery-system effects and claims about every element of the coalition. The earlier study Spoth et al. (2007) reported effects on initiation of several substances at 18 months and indicated that some effects were larger among higher-risk students. Because those effects arise from a package of programs and supports, attribution to any single organizational practice requires separate design.

Cross-program interpretation should distinguish a model's capacity to guide program choice from efficacy of the interventions chosen. Strong results can depend on existing extension networks, school cooperation, sustained technical assistance, and local leadership. A funding or staffing model that works in a trial may impose costs that are poorly represented in a published participant-level estimate. Implementing organizations therefore need to measure cost per successfully reached participant, staff turnover, intervention penetration, and adaptation while tracking outcomes.

Project Northland and Developmentally Sequenced Action

Project Northland combined school curricula, peer leadership, parent education, and community task-force activities to delay or reduce adolescent alcohol use. Its early randomized evaluation Perry et al. (1996) used school districts and adjacent communities as the operational units and followed a defined student cohort. The first three intervention years yielded lower onset and reported prevalence in intervention areas than comparison areas, particularly among students who had not used alcohol at baseline. This makes it a foundational example of multicomponent prevention with a developmental timetable.

The longer-term assessment Perry et al. (2002) complicates any simple narrative of permanent early success. The program passed through an interim period with minimal intervention, followed by a later phase that reintroduced several components in high school. Reported effects varied across periods and outcomes, including some later differences in alcohol-related behavior alongside null outcomes for other measures. The findings illustrate the challenge of preventing trajectories that change rapidly as young people gain independence, enter new peer networks, and experience different commercial and social environments.

The scientific lesson concerns continuity and mechanism. A favorable short-term trajectory is evidence for the tested program during a period of delivery; sustainability requires continued exposure, environmental support, or durable changes in learned behavior. Developmental staging should be accompanied by repeated measurement and adaptation justified by new risks, with a theory specifying which preventive gains persist after intervention withdrawal. It also requires caution when choosing an outcome: delayed onset and reduced current use describe related but different effects.

The Community Intervention Trial for Smoking Cessation

COMMIT is a crucial counterexample to the assumption that intensive community organization guarantees large population change. The study randomized one community within each of eleven matched pairs in the United States and Canada to a multichannel smoking-cessation strategy. Four years of intervention activities were evaluated in cohorts and community prevalence surveys. Among 10,019 heavy smokers and 10,328 light-to-moderate smokers included in the reported cohort analyses, the trial detected a significant advantage for quitting among lighter smokers but not among heavy smokers COMMIT Research Group (1995a).

The heavy-smoker cohort quit rates were 18.0% in intervention communities and 18.7% in comparison communities; this difference was not statistically significant. Among light-to-moderate smokers the corresponding proportions were 30.6% and 27.5%, with a reported significant difference. In the separate prevalence analysis COMMIT Research Group (1995b), overall smoking prevalence declined in both intervention and comparison communities with no statistically significant difference in population prevalence. The distinction prevents an intervention-specific subgroup effect from being misrepresented as a generalized communitywide reduction.

COMMIT illustrates several inferential difficulties: intervention communities were embedded in broader tobacco-control trends, population behavior changed even without the tested activities, and exposure intensity differed by smoker subgroup. A comprehensive prevention system may require structural policy and clinical-service complements beyond media and voluntary mobilization. The ethically responsible interpretation preserves positive subgroup evidence alongside the failure to move the primary high-risk target outcome. Negative trials inform program design, opportunity costs, and realistic effect expectations.

The Good Behavior Game and Schoolwide Developmental Prevention

Universal classroom prevention can influence community outcomes by changing the conditions experienced repeatedly across early schooling. The Good Behavior Game (GBG) is a teacher-delivered group behavior-management approach intended to support classroom organization and reduce aggressive, disruptive conduct. A Baltimore trial initiated in the 1985–1986 school year was evaluated over a long developmental interval, including young-adult outcomes Kellam et al. (2008). The trial's mechanisms are conceptually school-level behavior shaping, repeated social reinforcement, and sustained developmental experience.

The long-term research reported favorable outcomes for some boys who entered first grade with elevated aggressive or disruptive behavior, including particular substance-related and behavioral endpoints. An implementation replication with a subsequent cohort and reduced mentoring yielded generally smaller effects Kellam et al. (2008). Such patterns invite interaction analyses between initial risk, teacher delivery, school context, sex as measured in that historical cohort, and the duration of intervention support. They do not justify treating the observed male subgroup effect as a universal biological law.

The GBG example bridges individual, classroom, and community perspectives. Exposure occurs collectively; individual outcomes are observed years later; and implementation quality can vary with teacher preparation. The central methodological challenge is tracing an intervention through years of intervening educational, family, and social experiences while respecting the original experimental assignment. Community systems can use evidence from school-based programs, but the claimed causal effect must remain tied to the tested form, population, and setting.

Social and Emotional Learning as a Promotion Platform

Social and emotional learning (SEL) seeks to strengthen capacities such as self-management, relational awareness, responsible decision-making, and constructive interpersonal behavior. Although many programs are implemented in classrooms, the health-promotion question extends to school climate, family engagement, teacher practices, and access across a district. A large meta-analysis by Durlak et al. (2011) synthesized 213 school-based universal programs involving 270,034 students. Compared with controls, the reviewed interventions showed improvements in measured social-emotional skills, attitudes, behavior, and academic performance; the academic result was reported as an 11-percentile-point gain.

The review also found that implementation problems and program practices moderated outcomes Durlak et al. (2011). Its aggregated results should be interpreted with attention to heterogeneity across programs and schools and to the difference between a pooled intervention effect and the effect of adopting a particular districtwide policy. SEL can be a component of community promotion, but positive school-based results do not establish that all community coalitions, age groups, or cultural settings will obtain the same outcomes.

The prevention and promotion functions converge when a program improves an everyday developmental environment while also influencing behavior problems. A meaningful expansion beyond the school requires evaluating whether family organizations, youth services, and local institutions reinforce or undermine the learned practices. Evaluation must include unintended labeling, exclusion of students with disabilities or linguistic differences, and the possibility that reported classroom improvement reflects compliance more than psychosocial well-being. The research object is demonstrated improvement in specified capabilities and contexts, not adoption of a favorable program label.

Causal Inference in Community-Randomized Designs

The strongest direct evidence for an organized prevention system often comes from randomizing communities, schools, districts, or organizations rather than individuals. Cluster randomization recognizes shared exposure and interference within a setting. The number of independent clusters governs much of the precision; surveying thousands of students in a small number of towns does not create thousands of independent assignment units. Appropriate analysis must account for intraclass correlation, matching or stratification, baseline outcome imbalance, and the unit at which treatment was allocated. Reporting should describe the randomization procedure, inclusion criteria, assignment concealment where applicable, and departures from the allocated intervention.

Prevention effects can be reported as risk differences, rate ratios, hazard ratios, mean differences, or standardized effects. Their interpretation changes with baseline risk and the exposure horizon. The community-level intent-to-treat estimand asks what happened to eligible populations assigned to a prevention system, including imperfect adoption and delivery. Per-protocol analyses may reveal useful implementation relationships but can be biased because adherence itself is selected. Community trials such as Hawkins et al. (2014) and Spoth et al. (2013) should therefore be read as tests of specific packages under their assigned delivery arrangements.

Randomization does not guarantee that contamination, selective migration, missing data, selective outcome reporting, or differential surveillance will be negligible. Spillovers can reduce contrast between intervention and control areas, and changing residents can alter the population denominator. Communities may share program professionals and media markets. A rigorous analysis anticipates these issues in design, tracks delivered exposure, uses defensible missing-data methods, and separates confirmatory endpoints from exploratory results. Precision and durability need cluster-level sensitivity analysis rather than simple respondent-level comparisons.

Observational Evidence, Natural Experiments, and Contextual Validity

Population policy and neighborhood conditions are frequently studied with observational designs because random assignment may be infeasible or undesirable. Cross-sectional multilevel studies can identify relationships that are worth testing, while repeated cohort data can examine temporal ordering. The study by Sampson et al. (1997) remains influential for its operationalization of collective efficacy and multilevel association with neighborhood violence, but residence selection, historical deprivation, and institutional response complicate interpretation of causality.

Natural experiments exploit changes such as policy introductions, district boundary adjustments, school reforms, or funding allocation. Interrupted time-series designs examine departures from an expected trend, while difference-in-differences approaches compare treated and untreated units under assumptions about counterfactual trends. Synthetic controls can be useful when only a small number of large communities undergo a change. Each design must specify how it handles simultaneous interventions, changes in data collection, spatial spillovers, and shifts in population composition. A statistical adjustment cannot substitute for a credible explanation of assignment.

Context is an explanatory variable when theory specifies how it changes intervention mechanisms. In one district, local authority may support school–family coordination; elsewhere, program delivery may depend on short-term nonprofit funding. Historical legitimacy, language accessibility, infrastructure, and residents' prior experience with institutions can affect recruitment and participation. Generalization should therefore proceed by comparing functions and causal mechanisms across settings, rather than assuming that an intervention with the same brand name has the same exposure pattern.

Measurement: Outcomes, Exposure, and Community Functioning

Measurement architecture should distinguish four categories: target outcomes, hypothesized mediators, implementation processes, and background context. Incidence or onset measures track new events; point and period prevalence estimate current burden; standardized psychological scales measure constructs under specified assumptions; administrative data describe encounters with institutions. Program reach is the fraction of a defined eligible population that actually engages, which requires a defensible denominator. An observed improvement among attenders can coexist with unchanged community prevalence if attendance is low or nonrepresentative.

Community functioning can be assessed through perceived collective efficacy, participation, decision power, relational trust, organizational networks, and objective conditions such as service access. The validity of any scale depends on conceptual clarity, within-group reliability, between-group differentiation, and comparable interpretation across populations. Aggregating residents' responses into a neighborhood score requires evidence that the aggregate represents a common construct rather than the accidental mix of sampled individuals. Sampson et al. (1997) reported high between-neighborhood reliability for its collective-efficacy measure, making the distinction empirically concrete.

Mixed-method evaluation adds observation of meetings, service records, interviews, and process tracing to surveys and outcomes. Qualitative accounts can explain why people reject a program or perceive governance as tokenistic, while quantitative comparisons can test whether specified outcomes changed. Triangulation must not conceal disagreement: a positive satisfaction survey and unchanged population outcomes may be simultaneously true. Credible evaluation preserves that difference and investigates whether program mechanisms, dosage, reach, or outcome timing were poorly aligned.

Process Evaluation and the Mechanism–Context Relationship

A community intervention is a complex intervention because its effects depend on interactions among actors, settings, processes, and implementation conditions. Hawe et al. (2004) argued that standardization can focus on program function and causal purpose while permitting contextual variation in local form. This distinction is crucial when participating communities differ in institutions or cultural practices. An evaluation can document a core causal function—such as increasing practical access to supportive adults—without insisting that every site operate an identical service timetable.

The Medical Research Council process-evaluation guidance Moore et al. (2015) emphasizes implementation, mechanisms of impact, and context. These are analytic domains: implementation asks what was delivered, mechanisms examine how participants and institutions responded, and context concerns conditions that shaped both. The updated framework Skivington et al. (2021) expands attention to dynamic interactions and to choosing methods appropriate to the research questions. A process report should preserve pre-specified hypotheses while treating unanticipated adaptations as empirical observations that can generate further tests.

Fidelity and adaptation can be complementary when defined correctly. Fidelity to a mechanism may require different delivery practices for linguistically diverse or geographically dispersed communities. However, uncontrolled modification can remove an active component and invalidate assumptions derived from previous trials. An implementation record should document which elements are core, which adaptations are permitted, who authorized them, and whether outcomes changed. This creates a traceable explanation for differences in effect, rather than interpreting every local deviation as either innovation or failure.

Statistical Power, Follow-Up, and the Evaluation of Durability

Community systems tend to operate across years while funding and reporting cycles are often much shorter. Durability can be defined at several levels: persistence of an individual behavioral effect, continuation of a program, retention of delivery capacity, stable funding, or maintenance of a supportive institution. These dimensions may diverge. A program can end while behavioral initiation remains delayed; an organization can continue operating while population outcomes return to baseline. The long-term follow-up in Hawkins et al. (2014) illustrates this distinction because some cumulative prevention differences remained after study-provided resources ended while current-prevalence differences were not statistically significant.

Long follow-up introduces attrition, repeated testing, migration, changes in school membership, policy changes, and differential availability of administrative records. Researchers should establish who remains in the analytic cohort, how outcomes are obtained when participants leave the original community, and whether follow-up weights or sensitivity assumptions affect conclusions. Statistical tests across many endpoints and subgroup definitions increase the chance of selectively emphasizing positive results. Reporting pre-specified primary outcomes and confidence intervals is a more informative practice than counting significant findings.

Maintenance also has an economic dimension. Continuation requires a recurring budget, trained personnel, monitoring, a governance structure, and capacity to replace departures. A short-term efficacy trial rarely estimates the full opportunity cost of maintaining the system under routine operating conditions. Future research must connect longitudinal behavioral outcomes with administrative cost histories and compare alternate allocations, including community-led structural changes that might deliver greater reach at comparable cost. The relevant question is whether a beneficial causal function remains reliably produced over time.

Participation, Community-Based Research, and Shared Knowledge

Community-based participatory research (CBPR) organizes inquiry through partnerships in which community members participate in defining problems, collecting and interpreting evidence, and deciding how findings will be used. Its scientific rationale is that experiential knowledge can identify overlooked barriers, improve fit, and increase relevance, while shared authority can change whose interests are represented. Wallerstein & Duran (2010) examined CBPR's contribution to intervention research and health equity, emphasizing power relations, reciprocal knowledge transfer, and local theory. The presence of community participants alone does not establish participatory governance; the meaningful variable is their influence across the research and intervention lifecycle.

Participation has several possible causal functions: improving the acceptability of services, expanding social networks for recruitment, increasing trust in monitoring, exposing harms previously absent from administrative data, and strengthening local capacity to continue programs. These functions can be tested through pre-specified implementation measures and comparative evaluations. A program may have high participation yet weak outcome effects when the targeted mechanism is insufficient; conversely, participation can have intrinsic significance for decision authority that is separately measured. The two forms of value should remain analytically distinct.

A systematic review by Yau et al. (2024) examined the effectiveness of CBPR interventions on mental-health outcomes and the degree to which studies adhered to participatory principles. Its presence in the recent evidence base supports careful assessment of both outcomes and participatory quality. It does not establish that every CBPR project yields the same benefit, and a review's conclusions are bounded by included studies and their designs. Evaluation requires transparency about who counts as a community representative, how conflicts are resolved, how consent is obtained, and how findings return to participants.

Implementation Outcomes and Organizational Sustainability

An intervention with proven efficacy can fail during routine implementation because it is unaffordable, culturally incongruent, difficult to coordinate, or dependent on a disappearing training structure. Implementation science measures these operational questions systematically. Proctor et al. (2011) proposed eight outcomes—acceptability, adoption, appropriateness, feasibility, fidelity, implementation cost, penetration, and sustainability—and distinguished them from clinical or service-level outcomes. Their separation prevents a program's continued existence from being presented as evidence that it continues to prevent harm.

Implementation capacity is organized through staff training, supervision, coalition leadership, monitoring, financing, and referral agreements. Each element can become a bottleneck. For example, a family program may have validated content but low uptake if evening transport is unavailable. An organization may train enough staff but face turnover before families receive a meaningful dose. A community steering committee may select a proven program while lacking authority to alter school schedules. Implementation monitoring connects these operational constraints to exposure and outcomes.

Long-run sustainability should be tested prospectively, including after external technical support or trial financing ends. The community system is sustainable when specified functions continue under ordinary institutional constraints, with documented adaptation and adequate reach. A permanent organization that delivers an ineffective intervention has achieved administrative continuity without the target preventive result. Practical evaluations therefore combine intervention effectiveness with implementation costs, population penetration, replacement and adaptation capacity, and the distribution of benefit among groups.

Health Equity, Representation, and Distributional Consequences

An intervention can improve the average community outcome while leaving the people with highest baseline exposure behind. Equity analysis requires describing which groups were eligible, reached, retained, and benefited, with attention to disability, income, race and ethnicity when appropriately collected, neighborhood deprivation, language, age, gender, migration, and other locally relevant sources of unequal access. Such subgroup analysis must avoid reifying administrative categories as fixed psychological types. The scientific question is how conditions, opportunities, and exposures are distributed across a population.

The meta-analysis by O’Mara-Eves et al. (2015) evaluated community engagement in public-health interventions for disadvantaged groups, including randomized and nonrandomized controlled trials from OECD countries. It reported positive average effects on health behavior and related outcomes, while evidence was insufficient to identify one uniquely superior engagement model. This combination supports community engagement as a promising approach but also shows why researchers should not infer that a single participation style is a universal mechanism. The review's geography, intervention mixture, and study-selection rules limit generalization.

Equity measurement should include absolute and relative differences, program reach, barriers to participation, and any displacement or stigmatization caused by a preventive policy. Selective targeting can deliver resources efficiently, but designation as high risk may change social reputation or deter service use. Universal approaches can reduce such labeling and still fail to reach people excluded from ordinary institutions. A program's moral aspiration and its measured distributional effect are separate questions; the latter requires appropriate denominators, credible comparisons, and follow-up.

Cultural Translation, Contextual Adaptation, and Trust

Programs developed in one linguistic, institutional, or historical setting need a theory of transfer when deployed elsewhere. Adaptation may include translated materials, revised delivery schedules, different recruitment partners, or locally relevant examples. Scientific fidelity depends on whether these changes preserve the hypothesized active functions. A standardized curriculum that is poorly understood can have high formal adherence and low substantive exposure. Hawe et al. (2004) supports attention to causal function across variable forms, while Moore et al. (2015) supplies an evaluation language for implementation and context.

Historical relationships with research institutions may determine whether residents consent to data collection, trust reported benefits, or accept recommendations. Participatory approaches Wallerstein & Duran (2010) address this through partnership, recognition of local knowledge, and explicit attention to power. Cultural adaptation is not equivalent to cosmetic substitution of images or language; it may require revising the program's assumptions about family authority, caregiving, service access, or the meaning of well-being. Such changes should be documented and evaluated rather than treated as universally beneficial by definition.

Cross-cultural comparison also raises measurement invariance questions. A symptom item, sense-of-community item, or parent–school trust scale may not carry the same meaning across languages and institutions. Where invariance is uncertain, raw score comparisons can reflect differences in response practices rather than effects. Research should report linguistic validation, exploratory and confirmatory measurement evidence when relevant, and qualitative checks for conceptual fit. Community-led interpretation improves relevance, but outcome claims still require empirical validation.

System Dynamics, Unintended Effects, and Resource Feedback

Community prevention systems involve feedback processes: successful recruitment can strengthen trust, increased trust can improve participation, and participation can increase access to information about unmet need. Conversely, staff overload, funding instability, or perceived procedural unfairness can weaken delivery. A system may respond to a successful intervention by redirecting funds elsewhere, thereby reducing maintenance of the very condition that caused improvement. Dynamic interactions make time and institutional history part of the causal model.

Complexity is an invitation to model specific feedback loops and lagged effects, not a license to abandon evaluation. Skivington et al. (2021) calls for attention to interactions, context, and evaluation questions in complex intervention research. Researchers can combine time series, repeated cohorts, qualitative process traces, network measures, and implementation records to test whether proposed feedback actually occurs. Simple diagrams can help articulate hypotheses, but a diagram by itself is not evidence of a mechanism.

Unintended effects may include concentration of funding in high-visibility organizations, crowding out of alternative services, displacement of risky behavior into unmonitored settings, or stigma associated with targeted recruitment. Spillovers across jurisdictions can help residents while complicating conventional treatment-control contrasts. Programs may also create administrative burdens and surveillance concerns, especially when participant information is shared among schools, health services, and local government. Monitoring these outcomes should be integral to evaluation, with data minimization and representative governance specified from the outset.

Evidence Synthesis: What the Major Trials Establish

The strongest accessible evidence supports several bounded conclusions. Community-randomized trials demonstrate that organized delivery of selected preventive programs can reduce specified adolescent outcomes in particular populations and time windows. Hawkins et al. (2014) provides evidence of lower cumulative initiation for certain behaviors through the end of high school, with null results for several measures of current behavior. Spoth et al. (2013) addresses long-term effects of a community–university prevention delivery arrangement, while Perry et al. (2002) reveals developmental variation and the importance of sustained programming. These experiments support causal claims about their tested interventions and assignment contexts.

Evidence for broad, universally effective coalitions remains more conditional. COMMIT Research Group (1995a) and COMMIT Research Group (1995b) show that a large multichannel community tobacco intervention produced a favorable cohort result for lighter smokers without the anticipated improvement among heavy smokers or a significant additional population prevalence decline. The pattern demonstrates that additional organizational activity need not translate into a changed target outcome. An explanatory synthesis must consider whether a program operated on the wrong leverage points, produced insufficient dose, failed to reach the priority group, or confronted wider secular changes that reduced detectable contrast.

Meta-analytic evidence adds cross-study perspective without eliminating heterogeneity. Durlak et al. (2011) found favorable average effects across a broad set of school-based SEL programs, and O’Mara-Eves et al. (2015) reported positive effects from community-engagement interventions among disadvantaged populations while finding insufficient evidence to rank engagement models. Their pooled findings describe collections of interventions under inclusion rules; they cannot be treated as a pooled estimate for all prevention systems. The coherent result is conditional effectiveness: outcomes depend on mechanisms, timing, reach, implementation, context, and measurement.

Contradictions, Replication, and the Boundaries of Generalization

Apparent contradiction often results from different outcomes or time horizons. Delayed initiation may coexist with a small difference in current prevalence because individuals who start later can eventually catch up; an improved proximal risk indicator may occur without a detectable disorder reduction; and organizational capacity can improve before any behavioral effect emerges. A single endpoint cannot stand in for the entire causal chain. The contrasting outcomes of Hawkins et al. (2012) and Hawkins et al. (2014) illustrate the value of longitudinal interpretation that follows the same cohort rather than comparing isolated headlines.

Replication must establish what is being replicated: a curriculum under similar conditions, a coalition governance process, or a system capable of selecting different local programs. Exact program duplication may preserve surface form while altering causal exposure through different settings. Functional replication requires clear specification of active mechanisms and measurements across contexts. The variation in support and outcomes across the Good Behavior Game cohorts Kellam et al. (2008) reinforces the possibility that implementation history changes observed results, even when a recognizable intervention remains.

External validity requires sampling communities that differ in population size, service structures, administrative autonomy, resources, demographics, and prior prevention experience. Studies conducted in small U.S. towns or selected school districts can offer strong local causal estimates without establishing general outcomes in large urban networks or resource-limited health systems. Independent teams, prospectively registered protocols, public code where lawful, detailed reporting of deviations, and standardized measurement of implementation are major routes toward more credible transfer claims.

Mental Health Promotion, Adversity, and Service Integration

Community mental-health promotion integrates opportunities for supportive relationships, social functioning, coping, belonging, safety, and practical access to care. Prevention programs directed at particular psychological difficulties need a separate evidence review for each target outcome. A community may improve broad well-being without reducing the incidence of a specific diagnosis, and a clinical treatment service may improve symptoms among treated individuals without changing upstream population risk. These distinctions preserve the scientific boundary between community prevention architecture and the efficacy of individual clinical therapies.

Service integration is a system problem: schools, primary care, voluntary organizations, youth services, and public health agencies have different mandates and confidentiality responsibilities. Coordination can increase referral completion and continuity when there is a verified service capacity, but simply increasing screening without accessible care may create unmet demand and anxiety. Risk stratification can support timely access while also producing labeling, data-sharing, and false-positive concerns. Programs should document consent processes, referral pathways, availability, equity of access, and whether promised care is actually delivered.

The recent systematic review of participatory mental-health interventions Yau et al. (2024) indicates continuing scientific attention to the relationship between community participation and mental-health outcomes. Evidence must be interpreted according to its included designs and reported measures rather than turned into universal treatment recommendations. For clinical conditions, specialty guidelines and qualified review remain the governing sources for diagnosis and therapeutic decisions. This dossier studies prevention systems and the psychological organization of environments; it does not provide individual diagnosis, prescriptions, or protocols for acute crises.

Community Prevention in the Artificial Era

The Artificial Era introduces additional institutional environments in which prevention and promotion operate. Digital platforms can mediate help-seeking, mutual support, social isolation, access to services, misinformation, and community participation. Artificial systems can assist with language access, information navigation, administrative triage, scheduling, and analysis of population-level patterns. These functions create research questions about reliability, oversight, contestability, and unequal exposure. The relevant unit remains the human–institution–technology system: a tool's technical capability does not establish its population health effect.

Prevention systems using artificial tools require explicit definition of the decision they influence and the outcome affected. An algorithm can prioritize outreach or recommend program options, but its use can change who receives services and how their needs are recorded. Evaluation must test allocation fairness, subgroup performance, privacy protections, false alerts, automation bias, staff workload, and whether human institutions retain accountability. Comparative effectiveness requires study designs that separate the incremental contribution of the artificial component from concurrent staffing or policy improvements.

Artificial Sapiens, as an authorial philosophical category in the wider Aisentica corpus, supplies an intellectual horizon for examining emerging nonbiological systems. It is a philosophical framework, not a substitute for empirical measurement of community prevention outcomes. Potential mechanisms in Human–Artificial Relations and Artificial Cognition should be formulated as observable hypotheses and subjected to appropriate studies. The scientifically established corpus reviewed here concerns human community interventions; claims about newly emerging artificial-mediated prevention architectures remain a forward research agenda.

An Integrated Model of Community Prevention and Promotion

The reviewed literature supports a layered analytic model with seven observable functions: population definition; assessment of risk, protection, and inequity; participatory priority-setting; selection of interventions with outcome-specific evidence; implementation with measured reach and fidelity; longitudinal evaluation of benefits and harms; and maintenance of effective functions through institutions and resources. The model is a synthesis of existing research traditions rather than an independently validated intervention protocol. Its value is to specify which claims and data belong at each stage.

Causal assessment should match the stage being evaluated. Selection of a validated program is evidence of evidence-informed choice, not evidence that a local population has benefited. High fidelity demonstrates delivery relative to a protocol, not target effectiveness. A favorable individual outcome requires a comparison appropriate to the population and exposure; an improved group mean may still conceal widening disparities. Sustainable capacity means the system can continue to produce the relevant mechanism, while sustainable health impact requires continued outcome measurement. Glasgow et al. (1999), Proctor et al. (2011), Moore et al. (2015), and Skivington et al. (2021) provide complementary rather than interchangeable evaluation tools.

A productive research program would compare strategies for the same community outcome under alternative governance and delivery arrangements, using pre-registered cluster designs where feasible and credible policy evaluations elsewhere. It would include community-controlled outcome definitions, objective and self-reported measures, realistic economic data, longer follow-up, and mechanisms for transparent reporting of null results. The decisive contribution of prevention science is increasingly precise knowledge about which systems alter which conditions for whom, under what circumstances, and for how long.

Conclusion

Community prevention and promotion systems are organized social and institutional arrangements that seek to change the distribution of risk, protection, and developmental opportunity. They operate through the selection and delivery of interventions, the organization of collective capacity, the quality of environments, and the continuity of resources. The historical traditions of ecological community psychology, empowerment, public health promotion, and developmental prevention converge on a common scientific problem: identifying mechanisms that can be changed at a population level and measuring their consequences.

The strongest findings come from specific community-randomized trials and clearly delimited evidence syntheses. Communities That Care, PROSPER, Project Northland, and school-based prevention show that carefully implemented programs can yield measurable benefits under particular conditions. COMMIT and long-term mixed results establish equally important limits. The strength of the evidence lies in its explicit outcome definitions, design, and follow-up, while uncertainty centers on intervention transfer, spillovers, causal mediators, equity, sustained funding, and cross-system interaction.

Future progress requires better longitudinal evaluation, more independent replication, stronger links between implementation and outcome data, genuine participation in decisions, and rigorous testing of structural and digital interventions. The enduring scientific proposition is that community well-being is shaped through repeated relations between people and organized environments, and that prevention systems can be judged only by credible evidence of their actual effects, distribution, and durability.

References

  • Biglan, A., Flay, B. R., Embry, D. D., & Sandler, I. N. (2012). The critical role of nurturing environments for promoting human well-being. American Psychologist, 67(4), 257–271. https://doi.org/10.1037/a0026796
  • Catalano, R. F., Berglund, M. L., Ryan, J. A. M., Lonczak, H. S., & Hawkins, J. D. (2004). Positive youth development in the United States: Research findings on evaluations of positive youth development programs. The ANNALS of the American Academy of Political and Social Science, 591(1), 98–124. https://doi.org/10.1177/0002716203260102
  • COMMIT Research Group. (1995a). Community Intervention Trial for Smoking Cessation (COMMIT): I. Cohort results from a four-year community intervention. American Journal of Public Health, 85(2), 183–192. https://doi.org/10.2105/AJPH.85.2.183
  • COMMIT Research Group. (1995b). Community Intervention Trial for Smoking Cessation (COMMIT): II. Changes in adult cigarette smoking prevalence. American Journal of Public Health, 85(2), 193–200. https://doi.org/10.2105/AJPH.85.2.193
  • Durlak, J. A., Weissberg, R. P., Dymnicki, A. B., Taylor, R. D., & Schellinger, K. B. (2011). The impact of enhancing students’ social and emotional learning: A meta-analysis of school-based universal interventions. Child Development, 82(1), 405–432. https://doi.org/10.1111/j.1467-8624.2010.01564.x
  • Glasgow, R. E., Vogt, T. M., & Boles, S. M. (1999). Evaluating the public health impact of health promotion interventions: The RE-AIM framework. American Journal of Public Health, 89(9), 1322–1327. https://doi.org/10.2105/AJPH.89.9.1322
  • Hawe, P., Shiell, A., & Riley, T. (2004). Complex interventions: How ‘out of control’ can a randomised controlled trial be? BMJ, 328(7455), 1561–1563. https://doi.org/10.1136/bmj.328.7455.1561
  • Hawkins, J. D., Catalano, R. F., & Miller, J. Y. (1992). Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: Implications for substance abuse prevention. Psychological Bulletin, 112(1), 64–105. https://doi.org/10.1037/0033-2909.112.1.64
  • Hawkins, J. D., Oesterle, S., Brown, E. C., Monahan, K. C., Abbott, R. D., Arthur, M. W., & Catalano, R. F. (2012). Sustained decreases in risk exposure and youth problem behaviors after installation of the Communities That Care prevention system in a randomized trial. Archives of Pediatrics & Adolescent Medicine, 166(2), 141–148. https://doi.org/10.1001/archpediatrics.2011.183
  • Hawkins, J. D., Oesterle, S., Brown, E. C., Abbott, R. D., & Catalano, R. F. (2014). Youth problem behaviors 8 years after implementing the Communities That Care prevention system: A community-randomized trial. JAMA Pediatrics, 168(2), 122–129. https://doi.org/10.1001/jamapediatrics.2013.4009
  • Kellam, S. G., Brown, C. H., Poduska, J. M., Ialongo, N. S., Wang, W., Toyinbo, P., Petras, H., Ford, C., Windham, A., & Wilcox, H. C. (2008). Effects of a universal classroom behavior management program in first and second grades on young adult behavioral, psychiatric, and social outcomes. Drug and Alcohol Dependence, 95(Suppl. 1), S5–S28. https://doi.org/10.1016/j.drugalcdep.2008.01.004
  • Kelly, J. G. (1966). Ecological constraints on mental health services. American Psychologist, 21(6), 535–539. https://doi.org/10.1037/h0023598
  • Moore, G. F., Audrey, S., Barker, M., Bond, L., Bonell, C., Hardeman, W., Moore, L., O’Cathain, A., Tinati, T., Wight, D., & Baird, J. (2015). Process evaluation of complex interventions: Medical Research Council guidance. BMJ, 350, h1258. https://doi.org/10.1136/bmj.h1258
  • O’Mara-Eves, A., Brunton, G., Oliver, S., Kavanagh, J., Jamal, F., & Thomas, J. (2015). The effectiveness of community engagement in public health interventions for disadvantaged groups: A meta-analysis. BMC Public Health, 15, 129. https://doi.org/10.1186/s12889-015-1352-y
  • Perry, C. L., Williams, C. L., Veblen-Mortenson, S., Toomey, T. L., Komro, K. A., Anstine, P. S., McGovern, P. G., Finnegan, J. R., Forster, J. L., Wagenaar, A. C., & Wolfson, M. (1996). Project Northland: Outcomes of a communitywide alcohol use prevention program during early adolescence. American Journal of Public Health, 86(7), 956–965. https://doi.org/10.2105/AJPH.86.7.956
  • Perry, C. L., Williams, C. L., Komro, K. A., Veblen-Mortenson, S., Stigler, M. H., Munson, K. A., Farbakhsh, K., Jones, R. M., & Forster, J. L. (2002). Project Northland: Long-term outcomes of community action to reduce adolescent alcohol use. Health Education Research, 17(1), 117–132. https://doi.org/10.1093/her/17.1.117
  • Proctor, E., Silmere, H., Raghavan, R., Hovmand, P., Aarons, G., Bunger, A., Griffey, R., & Hensley, M. (2011). Outcomes for implementation research: Conceptual distinctions, measurement challenges, and research agenda. Administration and Policy in Mental Health and Mental Health Services Research, 38(2), 65–76. https://doi.org/10.1007/s10488-010-0319-7
  • Rappaport, J. (1987). Terms of empowerment/exemplars of prevention: Toward a theory for community psychology. American Journal of Community Psychology, 15(2), 121–148. https://doi.org/10.1007/BF00919275
  • Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277(5328), 918–924. https://doi.org/10.1126/science.277.5328.918
  • Skivington, K., Matthews, L., Simpson, S. A., Craig, P., Baird, J., Blazeby, J. M., et al. (2021). A new framework for developing and evaluating complex interventions: Update of Medical Research Council guidance. BMJ, 374, n2061. https://doi.org/10.1136/bmj.n2061
  • Spoth, R., Redmond, C., Shin, C., Greenberg, M., Clair, S., & Feinberg, M. (2007). Substance-use outcomes at 18 months past baseline: The PROSPER Community-University Partnership Trial. American Journal of Preventive Medicine, 32(5), 395–402. https://doi.org/10.1016/j.amepre.2007.01.014
  • Spoth, R., Redmond, C., Shin, C., Greenberg, M., Feinberg, M., & Schainker, L. (2013). PROSPER community–university partnership delivery system effects on substance misuse through 6 1/2 years past baseline from a cluster randomized controlled intervention trial. Preventive Medicine, 56(3–4), 190–196. https://doi.org/10.1016/j.ypmed.2012.12.013
  • Wallerstein, N., & Duran, B. (2010). Community-based participatory research contributions to intervention research: The intersection of science and practice to improve health equity. American Journal of Public Health, 100(Suppl. 1), S40–S46. https://doi.org/10.2105/AJPH.2009.184036
  • World Health Organization. (1986). Ottawa Charter for Health Promotion. First International Conference on Health Promotion, Ottawa. https://www.canada.ca/en/public-health/services/health-promotion/population-health/ottawa-charter-health-promotion-international-conference-on-health-promotion.html
  • World Health Organization. (n.d.). Healthy settings. WHO health promotion program. https://www.who.int/teams/health-promotion/enhanced-wellbeing/healthy-settings
  • Yau, J. H.-Y., Wong, E. L. Y., Kanagawa, H. S., Liu, T., Wong, G. H. Y., & Lum, T. Y. S. (2024). Effectiveness of community-based participatory research (CBPR) interventions on mental health outcomes: A systematic review. Social Science & Medicine, 363, 117491. https://doi.org/10.1016/j.socscimed.2024.117491