Research Paper Outline Generator

Research Paper Outline: Neighborhood Green Space Accessibility and Adolescent Depressive Symptoms

Target venue: Journal of Urban Health, Original Research Design: Explanatory sequential mixed methods (QUAN → qual) Style: APA 7th edition Word limit: 8,000. Confirm current JUH author guidelines for word, table, and abstract limits before submission, because journal limits may be stricter than 8,000.

Suggested Word Budget

JUH typically integrates the literature review into the Introduction. You may keep it as a distinct subsection under the Introduction and label it with APA Level 2 headings.

1. Working Title and Abstract Sketch

1.1 Working Title

• Primary option: Green Space Accessibility and Depressive Symptoms Among Adolescents in Three Mid-Sized U.S. Cities: An Explanatory Sequential Mixed-Methods Study
• Alternative (more findings-oriented, finalize after analysis): "It's There, But It's Not for Us": Objective and Perceived Green Space Access and Adolescent Depressive Symptoms in Fort Wayne, Spokane, and Chattanooga
• The title must:
  • name the population (adolescents), exposure (accessibility, not just "green space"), outcome (depressive symptoms, not "depression," since you use a screening measure), and design;
  • avoid causal verbs ("effect of," "impact of") given cross-sectional data.

1.2 Abstract Sketch (Structured: Background, Methods, Results, Conclusions)

• Aim: This study examined whether objective (GIS-derived) and perceived green space accessibility are associated with depressive symptom severity among adolescents aged 13–17, and how adolescents explain their use of, or avoidance of, neighborhood green space.
• Method: Cross-sectional survey data from approximately 600 adolescent–parent dyads in Fort Wayne (IN), Spokane (WA), and Chattanooga (TN), collected January 2023–June 2024, were linked to residential green space metrics and analyzed with multilevel regression. Twenty-five purposively sampled adolescents then completed semi-structured interviews, analyzed with reflexive thematic analysis and integrated through joint displays.
• Results placeholder: [Insert direction, magnitude with 95% CIs, and key effect modifiers; do not draft until the analysis is complete.]
• Contribution: The study extends evidence on green space and adolescent mental health beyond large metropolitan areas. It distinguishes availability from functional accessibility, and it uses qualitative data to explain heterogeneity that quantitative models alone cannot.

2. Introduction

2.1 Problem Framing

• Establish the burden of adolescent depressive symptoms in the U.S. and recent trends. [CITATION NEEDED: national surveillance data, e.g., YRBS or NSDUH major depressive episode prevalence among adolescents]
• Position the neighborhood built environment as a modifiable, population-level determinant, in contrast to individual-level clinical interventions. [CITATION NEEDED]
• Introduce green space as a candidate protective factor with established plausibility in adults and emerging evidence in youth. [CITATION NEEDED: recent systematic review/meta-analysis on green space and youth mental health]

2.2 Significance: Why Adolescents, and Why Mid-Sized Cities

• Adolescents: Adolescence is a period of increasing independent mobility and a developmental window for the onset of depression. Their spatial exposures differ from those of adults and young children. [CITATION NEEDED: developmental epidemiology of depression onset; adolescent independent mobility literature]
• Mid-sized cities: The evidence base is dominated by large metros (e.g., NYC, Los Angeles) and European cohorts. Mid-sized U.S. cities have distinct land-use patterns, including lower density, car dependence, and uneven park distribution, which may change how access operates. [CITATION NEEDED: evidence that U.S. green space research is concentrated in large metros; census definition of mid-sized city used]
• Site variation: Explain why these three cities were chosen. They differ in climate, region, and racial and economic composition, which gives analytic variation while holding city size roughly constant. State explicitly whether the sites were chosen by design or by convenience (existing partnerships). Reviewers will ask.

2.3 Research Questions

• RQ1 (QUAN): Is residential green space accessibility, measured (a) objectively and (b) as perceived by adolescents, associated with depressive symptom severity among adolescents aged 13–17, net of individual, household, and neighborhood confounders?
• RQ2 (QUAN, effect modification): Does this association differ by gender, race/ethnicity, household income, neighborhood deprivation, or city?
• RQ3 (qual → integration): How do adolescents describe their use of, barriers to, and emotional experiences in neighborhood green space? How do these accounts explain convergence or divergence between objective access and depressive symptoms?

2.4 Hypotheses

• H1: Greater objective and perceived green space accessibility will be associated with lower depressive symptom scores.
• H2: Perceived accessibility will show a stronger association with depressive symptoms than objective availability. This follows from the argument that use, not proximity, is the operative pathway. [CITATION NEEDED]
• H3 (equigenic hypothesis): Associations will be stronger among adolescents in higher-deprivation neighborhoods. [CITATION NEEDED: equigenesis literature]
• Pre-specify the direction of each hypothesis. State whether they were pre-registered (e.g., OSF). If they were not, label analyses as confirmatory or exploratory in the Methods.

3. Literature Review

Organize the review into four thematic clusters that end in a clearly stated gap.

3.1 Cluster A: Theoretical Pathways Linking Green Space and Mental Health

• Summarize the major frameworks: Attention Restoration Theory, Stress Reduction Theory, and the three-domain pathway model (reducing harm, restoring capacities, building capacities). [CITATION NEEDED for each]
• Identify which pathways are most plausible for adolescents: physical activity, social cohesion and peer contact, stress and rumination reduction, and sleep. [CITATION NEEDED]
• Explain how the framework justifies your covariate and mediator choices. This section should produce a directed acyclic graph (DAG), which will appear as a figure in the Methods.

3.2 Cluster B: Empirical Evidence in Children and Adolescents

• Synthesize longitudinal and cross-sectional studies of youth depressive or internalizing symptoms. Note that the evidence is mixed and that effect sizes are generally small. [CITATION NEEDED: systematic reviews; key cohort studies, e.g., Danish registry and UK/European cohorts]
• Separate studies of early childhood from studies of adolescence, since exposure mechanisms differ.
• Highlight the geographic skew toward Europe and large metropolitan areas. [CITATION NEEDED]

3.3 Cluster C: Measurement of Green Space Exposure

• Contrast the main exposure types:
  • greenness, such as NDVI within buffers;
  • availability, such as park area within a buffer;
  • accessibility, such as network distance to the nearest park entrance;
  • quality and perceived access. [CITATION NEEDED]
• Discuss the modifiable areal unit problem (MAUP) and the uncertain geographic context problem (UGCoP). Residential buffers may misrepresent adolescent activity spaces, which include school, friends' homes, and transit routes. [CITATION NEEDED]
• Note weak correlations between objective and perceived measures in prior work, which justifies measuring both. [CITATION NEEDED]

3.4 Cluster D: Equity, Safety, and Social Meaning of Green Space

• Review evidence that park quality, safety, policing, and racialized experiences of public space shape youth use. [CITATION NEEDED]
• Discuss green gentrification and differential benefits by neighborhood SES and race. [CITATION NEEDED]
• Note the scarcity of qualitative work capturing adolescents' own accounts in U.S. settings. [CITATION NEEDED]

3.5 Knowledge Gap Statement

• State the gap in a single, specific sentence, for example: "No U.S. study has combined GIS-derived and perceived green space accessibility with adolescent-reported depressive symptoms in mid-sized cities, nor used qualitative data to explain why proximity does or does not translate into mental health benefit." Verify this claim through a documented search before asserting it. [CITATION NEEDED / verify through systematic search]
• Connect each part of the gap directly to RQ1–RQ3.

4. Methodology

4.1 Research Design

• Explanatory sequential mixed-methods design (QUAN → qual): The quantitative phase answers RQ1 and RQ2. Its results drive purposive sampling and the interview guide for the qualitative phase (RQ3). [CITATION NEEDED: Creswell & Plano Clark mixed-methods design typology]
• Justify the design: surveys estimate associations but cannot explain residual heterogeneity or mechanisms, and interviews supply explanatory depth.
• Name the integration points: sampling (connecting), data collection (building), and interpretation (merging via joint displays). [CITATION NEEDED: Fetters et al. on integration]
• Reporting standards: STROBE (cross-sectional), COREQ or SRQR (qualitative), GRAMMS or APA JARS-Mixed. Include completed checklists as supplementary files.

4.2 Setting and Sampling

4.2.1 Quantitative Sample

• Specify the recruitment frame and method (e.g., school-based, address-based sampling, community partners, or panel). Report response rates by city. [AUTHOR MUST SUPPLY]
• Report inclusion and exclusion criteria: age 13–17; residence at the current address for ≥ X months, which reduces exposure misclassification; English or Spanish fluency (if applicable); and one adolescent per household.
• Include a participant flow diagram (STROBE): contacted → eligible → consented → complete dyads → geocoded → analytic sample.
• Power or sensitivity analysis: Report the minimum detectable standardized effect at n≈600 with α = .05 and 80% power, accounting for the design effect from neighborhood clustering (use the ICC from pilot data or the literature). [AUTHOR MUST COMPUTE] Acknowledge that interaction tests (RQ2) will be underpowered and frame them as exploratory.

4.2.2 Qualitative Sample (n = 25)

• Purposive maximum-variation sampling from quantitative participants who consented to recontact. Stratify on the exposure × outcome cross-classification:
  • concordant cases (high access–low symptoms; low access–high symptoms);
  • discordant cases (high access–high symptoms; low access–low symptoms), which are the most analytically valuable;
  • balance across city, gender, and race/ethnicity.
• Justify n = 25 with an information power argument or evidence on meaning saturation, rather than asserting "saturation was reached." [CITATION NEEDED: Malterud et al.; Hennink et al.]
• Report the timing gap between survey and interview, since longer gaps allow symptom states to change.

4.3 Data Sources and Measures

4.3.1 Outcome: Depressive Symptoms

• Use a validated adolescent self-report instrument, such as the PHQ-A, the Short Mood and Feelings Questionnaire (SMFQ), or the CES-D. Report the scale, score range, and internal consistency (Cronbach's α or McDonald's ω) in this sample. [CITATION NEEDED: validation study for chosen instrument in U.S. adolescents]
• Treat the outcome primarily as continuous to preserve power. Run a secondary analysis at a clinical cutoff, and justify the chosen threshold with a citation.
• If parent-report of adolescent symptoms was collected, use it in a sensitivity analysis. Discuss the expected informant discrepancy. [CITATION NEEDED]
• Ethics: Describe the safety protocol for suicidal ideation items (e.g., PHQ-A item 9), including referral procedures.

4.3.2 Exposure: Green Space Accessibility (Multiple Operationalizations)

• Greenness: mean NDVI within 250 m, 500 m, and 1,000 m street-network or Euclidean buffers around the geocoded residence. Use cloud-free imagery (e.g., Landsat or Sentinel-2) from the growing season. Justify the season choice, because Spokane, Fort Wayne, and Chattanooga differ in phenology.
• Park accessibility: network distance to the nearest public park entrance (not the centroid), and park acreage within a 10-minute walk. Specify the data source (e.g., municipal parcel data, Trust for Public Land ParkServe) and vintage. [AUTHOR MUST SUPPLY]
• Perceived access and quality: a validated or adapted adolescent-report scale covering perceived availability, safety, attractiveness, and frequency of use. Report its psychometrics. [CITATION NEEDED]
• Pre-specify a primary exposure (recommended: network distance or perceived access, depending on the hypothesis) to avoid multiplicity across 5+ exposure metrics. Treat the other measures as secondary. Apply a false discovery rate correction or report them transparently as exploratory.

4.3.3 Covariates (DAG-Informed)

• Individual: age, gender identity, sex assigned at birth (if collected), race/ethnicity, pubertal development, and sexual orientation (if collected; a known depression risk factor).
• Household (from parent report): income-to-poverty ratio, parental education, parental depressive symptoms (a major confounder, and a strength of the dyadic data), household structure, and length of residence.
• Neighborhood: Area Deprivation Index or Social Vulnerability Index, population density, perceived and/or police-recorded crime, and urbanicity within the city.
• Temporal: season and month of survey completion, since data span 18 months and depressive symptoms may vary seasonally.
• Do not adjust for mediators (physical activity, social cohesion, sleep) in the primary model. Examine them separately (see 4.4.4).

4.3.4 Qualitative Data

• Semi-structured interview guide, developed from quantitative results, covering: routine spaces, green space use and non-use, safety, social meaning, emotional experiences, and COVID-era changes (if relevant).
• Report mode (in-person or virtual), duration, interviewer characteristics, audio recording, transcription, and compensation.
• Consider a participatory element, such as participant-generated photos or map annotation, to address activity-space limitations.

4.4 Analysis Plan

4.4.1 Descriptive and Preliminary Analyses

• Report sample characteristics by city and compare respondents with non-respondents (if available).
• Report correlations between objective and perceived exposure measures.
• Characterize missing data patterns (MCAR test or patterns table).

4.4.2 Primary Models (RQ1)

• Multilevel linear regression with adolescents nested in census tracts or block groups, using random intercepts. Report the ICC. Use successive models: unadjusted, then adjusted for individual and household covariates, then additionally adjusted for neighborhood covariates.
• Include city as a fixed effect, not a random effect. Three clusters are too few to estimate a variance component reliably. [CITATION NEEDED: literature on minimum number of higher-level units]
• Check for spatial autocorrelation of residuals (Moran's I). If it is present, use spatial error or lag specifications as a sensitivity analysis.
• Standardize exposures (per IQR or per SD) for interpretability. Report βs with 95% CIs, not only p-values.

4.4.3 Effect Modification (RQ2)

• Test pre-specified interaction terms (exposure × gender, × race/ethnicity, × neighborhood deprivation, × city). Present stratified estimates.
• Label these tests exploratory given power limitations. Avoid interpreting non-significant interactions as evidence of no difference.

4.4.4 Secondary: Pathway Exploration

• Conduct exploratory mediation through physical activity, perceived social cohesion, and sleep, using counterfactual-based methods.
• State the key caveat explicitly: cross-sectional mediation cannot establish temporal ordering, so present findings as hypothesis-generating only. [CITATION NEEDED: cautions on cross-sectional mediation]

4.4.5 Missing Data and Sensitivity Analyses

• Use multiple imputation by chained equations under the MAR assumption, with the number of imputations justified by the fraction of missing information. Compare with complete-case results.
• Sensitivity analyses:
  • alternative buffer sizes (addresses MAUP);
  • restriction to adolescents with ≥ 2 years at the current address;
  • clinical cutoff outcome;
  • parent-reported outcome;
  • E-values for unmeasured confounding. [CITATION NEEDED: VanderWeele & Ding]

4.4.6 Qualitative Analysis

• Use reflexive thematic analysis, with an inductive–deductive hybrid informed by the pathway framework. [CITATION NEEDED: Braun & Clarke]
• Describe the procedures: two coders, codebook development, analytic memos, team debriefing, and a reflexivity statement on researcher positionality. For reflexive TA, avoid framing inter-rater reliability (kappa) as the quality criterion, because it conflicts with that epistemology.
• Name the software (e.g., NVivo, ATLAS.ti, Dedoose).

4.4.7 Mixed-Methods Integration

• Build joint displays that align quantitative subgroups (e.g., discordant cases) with qualitative themes. [CITATION NEEDED]
• Classify findings as convergence, complementarity, expansion, or discordance, and state explicitly how discordance will be handled.
• Produce meta-inferences that answer RQ3 and speak back to RQ1 and RQ2.

4.5 Ethics

• Report IRB approval (institution and protocol number), parental permission and adolescent assent, and confidentiality provisions for minors.
• Describe the handling of residential geocodes (secure storage, geomasking in any published maps) and the distress and referral protocol.

4.6 Design Limitations (Preview; Full Discussion in Section 6)

• The cross-sectional design precludes causal inference and leaves reverse causation possible: depressed adolescents may use green space less or perceive it less favorably, which inflates associations with perceived measures.
• Residential self-selection: families who value green space may differ systematically.
• Generalizability is limited to three non-random cities.
• Common-method bias affects perceived exposure and self-reported outcome.

5. Expected Results / Analysis

Do not report hypothesized numbers. This section of the outline specifies what patterns would support or refute each hypothesis and how each will be presented. Replace it entirely with actual results.

5.1 RQ1: Main Association

• Supportive pattern for H1: Adjusted estimates show lower depressive symptom scores with greater accessibility, and the direction is consistent across buffer sizes.
• Supportive pattern for H2: Estimates for perceived access are larger in standardized terms than those for NDVI or distance. Interpret with caution, because shared-method variance could produce the same pattern.
• Null or attenuated results: Plan the interpretation now. Distinguish a precise null (narrow CI around zero) from an inconclusive estimate (wide CI). Pre-commit to reporting null findings, since publication bias in this literature is a concern. [CITATION NEEDED]
• Presentation: Table 2 shows nested models. A forest plot compares exposure metrics.

5.2 RQ2: Heterogeneity

• Supportive pattern for H3: Larger protective associations in high-deprivation tracts.
• Anticipated gender differences: Report them, but consider both directions plausible. Prior evidence is inconsistent, and adolescent girls may face distinct safety barriers. [CITATION NEEDED]
• City differences: May reflect climate, park system design, or unmeasured confounding. Interpret descriptively.
• Presentation: Stratified forest plot and interaction p-values with FDR-adjusted q-values.

5.3 RQ3: Qualitative Themes and Integration

• Anticipated thematic domains: These sensitizing concepts, not predetermined findings, should remain open to revision:
  • safety and surveillance, including racialized experiences;
  • social function (peer spaces versus isolation);
  • restorative use and emotional regulation;
  • practical barriers (transportation, time, amenities not designed for teens).
• Discordant cases: These are the key explanatory test. For example, high objective access with high symptoms may be explained by perceived unsafety or a lack of teen-appropriate amenities, and low access with low symptoms by alternative restorative spaces such as school grounds or rural relatives.
• Presentation: A joint display (Table 4) crossing exposure-outcome quadrants with themes and illustrative quotations. A quotation table with participant descriptors (pseudonym, age, city), not identifying details.

6. Discussion and Implications

6.1 Summary of Principal Findings

• Answer each RQ in one or two sentences, in order, without repeating statistics.
• State which hypotheses were supported, partially supported, or not supported.

6.2 Interpretation in Relation to Prior Evidence

• Compare effect direction and magnitude with prior youth studies and meta-analyses. [CITATION NEEDED]
• Use qualitative findings to explain divergence from prior work, which is the core mixed-methods value-add.
• Reconsider theoretical pathways: which were supported and which were challenged by adolescent accounts.

6.3 Theoretical Contributions

• Refine the availability → access → use → benefit chain with adolescent-specific evidence on where it breaks.
• Assess the equigenic hypothesis in U.S. mid-sized city contexts.
• Make a methodological contribution: demonstrate the added value of objective-perceived measurement and discordant-case sampling.

6.4 Practical and Policy Implications

• Urban planning and parks: teen-oriented park design, programming, lighting, and safe routes. Keep recommendations proportional to the strength of the cross-sectional evidence.
• Public health practice: integration with school-based mental health efforts and nature prescription programs for youth, with caution about overpromising. [CITATION NEEDED]
• Equity: anti-displacement considerations when greening low-income neighborhoods. [CITATION NEEDED: green gentrification]
• Tailor recommendations to mid-sized city governance and budget realities.

6.5 Strengths and Limitations

• Strengths:
  • multi-city design;
  • dyadic data allowing adjustment for parental depression;
  • multiple exposure operationalizations;
  • integration of adolescent voice.
• Limitations:
  • cross-sectional design and reverse causation;
  • residential self-selection;
  • residence-based exposure (UGCoP);
  • NDVI's insensitivity to quality and usability;
  • self-report and common-method bias;
  • 18-month data collection window (seasonal and secular trends);
  • limited generalizability;
  • qualitative sample drawn from consenting responders.
• For each major limitation, state the likely direction of bias where possible.

6.6 Future Research Directions

• Longitudinal cohorts with repeated exposure and outcome measurement.
• GPS or ecological momentary assessment to capture actual activity-space exposure and momentary mood.
• Natural experiments, such as new park openings or greening interventions, using difference-in-differences designs. [CITATION NEEDED: examples of greening natural experiments]
• Youth participatory action research to co-design park interventions.

6.7 Conclusion

• In three or four sentences, give the core answer, its importance for adolescent mental health in mid-sized cities, and one actionable takeaway.

7. References: Source Types and Search Strategy

No citations are listed below. Retrieve, read, and verify every source before citing it, and format in APA 7th edition.

7.1 Representative Source Types to Locate

1. Systematic reviews and meta-analyses on green space and child or adolescent mental health, internalizing symptoms, or depression.
2. Theoretical foundations: Attention Restoration Theory, Stress Reduction Theory, and the multi-pathway green space–health framework (reducing harm, restoring capacities, building capacities).
3. Large longitudinal cohort studies of childhood green space exposure and later psychiatric outcomes (e.g., European registry-based and birth cohorts).
4. U.S.-based studies of adolescent neighborhood environments and mental health, especially outside major metropolitan areas.
5. Exposure measurement methods papers: NDVI validity, buffer selection, network-based park accessibility, MAUP, and UGCoP.
6. Psychometric validation studies for your chosen depression instrument (PHQ-A, SMFQ, or CES-D) and your perceived-access scale in adolescent samples.
7. Mixed-methods methodology: design typologies, integration strategies, joint displays, and GRAMMS or JARS-Mixed reporting standards.
8. Qualitative methodology: reflexive thematic analysis, information power, and meaning saturation.
9. Environmental justice and equity literature: park equity, green gentrification, and racialized experiences of public space among youth.
10. Causal inference methods: DAGs in environmental epidemiology, E-values, and cautions on cross-sectional mediation.
11. Surveillance data sources: CDC YRBS or NSDUH reports for adolescent depression prevalence (grey literature).

7.2 Databases

• PubMed/MEDLINE, PsycINFO, Web of Science, Scopus, GreenFILE, and CINAHL.
• Grey literature: CDC, Trust for Public Land reports, and municipal park master plans for the three study cities.

7.3 Search Term Strings (Adapt Syntax per Database)

• ("green space" OR greenspace OR "green space" OR "urban nature" OR park* OR NDVI OR greenness)
• AND (adolescen* OR teen* OR youth OR "young people")
• AND (depress* OR "depressive symptoms" OR "internalizing" OR "mental health")
• Secondary strings:
  • "park access*" AND equity
  • "uncertain geographic context problem"
  • "joint display" AND "mixed methods"
  • "equigenic" OR "equigenesis"
  • "mid-sized cit*" OR "small and medium-sized cities" AND "built environment"
• Keep a search log (dates, databases, hits). This makes the "no prior study" gap claim defensible to reviewers.

Reviewer Checklist

An examiner or JUH reviewer is likely to ask:

1. Causality and reverse causation: Given the cross-sectional design, how have you ruled out the possibility that depressed adolescents simply perceive or use green space less? Does the framing avoid causal language throughout, including the title, abstract, and implications?
2. Exposure definition and multiplicity: Which green space metric is primary, and was it pre-specified? How did you handle multiple testing across buffers and metrics? Does a residence-based buffer meaningfully capture adolescent exposure?
3. Confounding and clustering: Is the covariate set justified by a DAG? Are parental depression, neighborhood deprivation, and residential self-selection addressed? How is clustering handled with only three cities?
4. Mixed-methods integration: Is integration genuine, through sampling from discordant cases and joint displays that produce meta-inferences? Or are the quantitative and qualitative strands merely reported side by side? Is n = 25 justified by a stated rationale rather than a bare saturation claim?
5. Generalizability and contribution: Why these three cities, and to which populations can the findings reasonably transfer? Does the stated knowledge gap hold up against a documented literature search?